(1.1)什么是 r&b?

r&b的全名是 rhythm & blues，一般译作“节奏怨曲”或“节奏布鲁斯”。 广义上， r&b可视为“黑人的流行音乐”，它源於黑人的 blues音乐，是现今西行流行乐和摇滚乐的基础， billboard杂志曾介定 r&b为，所有黑人音乐除了 jazz和 blues之外，都可列作 r&b，可见 r&b的范围是多么的广泛。 近年黑人音乐圈大为盛行的 hip hop和 rap都源於 r&b，并且同时保存着不少 r&b成分。

(1.2)什么是house?

house是於八十年代沿自 disco发展出来的跳舞音乐。 这是芝加哥的dj玩出的音乐， 他们将德国电子乐团kraftwerk的一张唱片和电子鼓(drum machine)规律的节奏 及黑人蓝调歌声混音在一,house就产生啦~ 一般翻译为”浩室”舞曲,为电子舞曲最基本的型式,4/4拍的节奏, 一拍一个鼓声,配上简单的旋律,常有高亢的女声歌唱. disco流行后，一些dj将它改变，有心将disco变得较为不商业化， bass和鼓变得更深沈， 很多时变成了纯音乐作品，即使有歌唱部分也多数是由跳舞女歌手唱的简短句子，往往没有明确歌词。 渐渐的，有人加入了latin(拉丁)、 reggae(瑞格源在西印度群岛)、 rap(说唱)或 jazz(爵士)等元素，至八十年代后期， house冲出地下范围，成为芝加哥、纽约及伦敦流行榜的宠儿。为什么会叫”house”呢?就是说只要你有简单的录音设备,在家里都做得出这种音乐~house也是电子乐中 最容易被大家所接受的.

cher唱的believe 就是个好例子.而m-people可说是house代表团体.house舞曲在1986年开始流行后,可说是取代了disco音 乐.

house可分为：

acid house 也就是融合了tb-303的acid声的house乐.deep house 有着相当浓厚的灵魂唱腔,又叫作garage,蛮流行化的. 像real mecoy, m-people, la bouche都是garage团体.hard house 简单来说,就是节奏较重,较猛的house.progressive house 这类house没啥灵魂唱腔,反而比较注重旋律和乐曲编排.有一 点像”演奏类”的house乐. 像sash!的专辑it’s my life便是很好的progressive house专辑.epic house 就是”史诗”house. epic house有着优美,流畅的旋律和磅礴的气势, 很少会有vocal在里面(几乎是没有!). bt的音乐就是很棒的epic house,而他也被称作”epic house天皇”. 其实连rock也有”epic rock”,trible house 想像一下:带有非洲原始风貌或是印地安人的鼓奏的house是啥样? 这就是trible house~ 这种house除了有一般house稳定的节奏外,在每拍之间,会加入一些带有原始风貌,零碎的鼓点. 蛮有趣的~ 不过,house的范围太广了,大家也不用硬要把一首曲子分类..这些只是告诉大家,house很多种”而已。

到了九十年代， house已减少了那前卫、潮流色彩，但仍是很受欢迎的音乐。

(2.1)什么是britpop?

britpop虽有个“ pop”字，但其实是 rock的一种，源於九十年代英伦，中文可译为“英式摇滚”，这是英伦乐坛对美国 grunge潮的一个回应，主要是以乐队形式出现。不过， britpop风格其实十分广泛，如 oasis是结他摇滚乐队， blur则 pop很多，而pulp则接近 glam rook及跳舞风格，不过他们都被列作 britpop。英国 britpop代表人物： oasis 、blur 、suede 、pulp 、radiohead

(2.2)什么是 trip-hop?

trip-hop是英伦/欧洲跳舞音乐的一种，它的名字来源是“ trip+ hip-hop= trip-hop”,因为它发源自英国的bristol，因此最早时称作“bristol hip-hop”。由于把把hip-hop(其实很多音乐都是架构在hip-hop上的……不知啥是hip-hop的去看看跳街 舞的人， 他们多半是用hip-hop音乐来跳的。)节奏变慢(有时很慢很慢)加入一些迷幻的味道，如很阴沉，肥厚的bass，轻微但迷幻 的合成音效，或是些唱片的取样，有时可能连唱片的杂音都会被“故意”取样进去。所谓“ trip”，指迷幻，氤氲的药物【旅程】，所以， trip-hop是种慢板的迷幻的、有 jazz感觉的、迷糊的、带点 hip-hop节奏的 break beat音乐。它虽然隶属跳舞音乐类，但其迷幻迷糊特色已令它跳一般跳舞音乐所有的明确节拍特色相去很远。

外国 trip-hop代表人物： trip-hop中公认的铁三角ortishead、 tricky、 massive attack 、mandalay moloko 、coldcut

(2.3)什么是 gangsta rap?

gangsta rap是 rap的一种，以 rap的内容多与都市罪案有关，充满暴力、色欲感受，这是反映现实的一种音乐路向。 gangsta rap於八十年代末期在美国兴起，音乐rap中的强悍尖锐派，在美国大受欢迎，唱片销路甚高。而不少 gangsta rap乐手本身真正“参与”现实中各式罪案，部分更因而入狱甚至死亡，可说是真正反映现实兼令人触目惊心的乐种。gangsta rap化表组合： ice-t、 ice cube、 dr. dre、 snoop doggy dogg、 2 pac、geto boys

(2.4)什么是 synth pop?

synth pop中的“ synth”，即 synthesizer，顾名思义， synth pop就是“由 sythesizrs炮制出来的流行乐”，当然除 synthesizers外还会用上其他电子乐器如电脑及鼓机等等。 synth pop於八十年代初期开始流行，至八十年代中开始沉寂，当年在香港也曾掀起过一阵热潮。synth pop的特点是科技感强，有时会颇冰冷，歌曲多是“三分钟流行曲”(3-minutes pop)，很多时 synth pop乐手会作入时打扮。synth pop代表组合： depeche mode、 human league、 duran duran、 omd、 gary numan

(2.5)什么是 orchestra ?

orchestra就是管弦乐团，它分成四部分： 1弦乐，包括小提琴、大提琴等， 2铜管乐 3木管乐和 4敲击乐四组。弦乐组每种乐器有多人演奏(竖琴除外)，四组演奏者由一人统筹兼领导，他就是乐团的指挥。orchestra於1 7世纪出现，到18世纪因海顿和莫扎特的作品而清楚地建立模式。19世纪加入了些新乐器，乐团人数加大。 orchestra是西方古典/正统音乐的正宗。西方流行/摇滚乐也经常运用 orchestra的部分或全部团员协助演出。

(2.6)什么是 chamber pop?

chamber pop是指典雅、高贵、精致的一种流行乐，它有一定的古典音乐感觉。 chamber music一词来自古典音乐，中文叫“室内乐”，是种小组弦乐演奏曲式，气氛高雅。chamber pop於九十年代兴起，是对当时的 lo-fi及 grunge的一种反应，强调优美的旋律、精致的配乐、乾净的录音，每每多用弦乐、管乐制造巴洛克时代的音乐感觉。外国代表乐队有the divine comedy， rialto， eric matthews及 balle & sebastian等，香港则有黄耀明、部分的彭羚、部分的 multiplex等等。

(2.7)什么是民歌( folk)?

民歌( folk) 原本是指每个民族的传统歌曲，每个民族的先民都有他们自原始/古代已有的歌曲，这些歌绝大部分都不知道谁是作者，而以口头传播， 一传十十传百，一代传一代的传下去至今。不过今天我们所说的民歌( folk)，大都是指流行曲年代的民歌( folk

)，所指的是主要以木结他为伴奏乐器，以自然坦率方式歌唱，唱出大家纯朴生活感受的那种歌曲。美国民歌手 woody guthrie在五十年代的唱片可说是最早的民歌唱片录音，所以普遍被认定是现代民歌( folk)的祖师。之後 pete seeger、 the weavers继续推动这

类音乐，六十年代越战，反战民歌手如 bob dylan、joan baez、peter、paul and mary等成为时代的呼声。後民歌向 pop、rock及都市化发展， bob dylan发明了 folkrock、 simon & garfunkel发展出中产口味的都市 folk pop，风行一时。八十年代 suzanne vega、

tracychapman等走出一种更富现在都市感觉的 urban folk(城市)/ contemporary(当代) folk路线。民歌 (folk)在英国、香港等乐坛也发展出不同的面貌。民歌(folk)近年较新的发展是与 new age结合(如 enya)，及与 trip hop结合)如 beth orton)。

(2.8)什么是 bossa nova?

bossa nova是种带 jazz味道的巴西音乐， 1950年代作曲家 antonio carlos jobim将巴西音乐节奏与美国西岸 cool jazz混合而成，柔和、舒服、轻松、懒洋洋、浪漫乃特色。每两个 bar的第1、4、7、11、14拍为重拍。歌手 joao gilberto的纯厚歌声是 jobim音乐的

出色拍档， 50年代中期传入美国後大受欢迎，其他着名 bossa nova乐手有 charlie byrd,stan getz和 astrud gilberto等，最着名的歌曲为“the girlfrom ipanema”。 ossa nova於60年代最流行，其馀年代亦不断有捧场者。

(2.9)什么是 classical pop ?

classical pop是指带古典响乐味道的流行曲，多用弦乐伴奏的歌曲都可列入此类，例如the beatles的“eleanorrigby”就是。澳洲乐队 crowded house的一些作品也可列入此类。如果说 classic pop，则是指经典的流行曲。

(2.10)什么是acappella ?

acappella是指没有乐器伴奏的歌曲，但凡纯以人声唱的歌都是 acappella，不过今天我们说 acappella通常是指有多重和唱的那种唱法，连乐器伴奏都由人声唱出。 acappella的相反是 instrumental，即纯音乐乐曲，任何类型的歌曲都可以以 acappella形式唱出。香港人最熟悉的 acappella歌曲是“so much in love”，香港 acappella乐队有姬声雅士等。

(3.1)什么是 world music?

world music是西方角度观点的词汇，意思指非英、美及西方民歌/流行曲的音乐，通常指发展中地区或落後地区的传统音乐，例如非洲及 南亚洲地区的音乐，有些地区如拉丁美洲的音乐，则能普及到自成一种类型。今天大家说的 world music通常是指与西方音乐混和了风格的、改良了的传统地区音乐。非洲的 king sunny ade，东欧的don byron，中国的朱哲琴，巴基斯坦的 nusrat fatehali kahn等是西方乐迷较熟悉的 world music乐手。

(3.2)什么是 dream-pop?

dream-pop是种“梦”般的流行曲，它有一种迷离的气氛，多靠synthesizers(电子合成器)造成，加了echo 效果的电结他也是重要的成分，歌唱部分往往很 “breathy”即呼吸声重，歌词也往往有梦般的诗意色彩。代表乐队有 cocteau wins、 lisa germano、 st. e

tienne、 this mortal coil、 my bloody valentine、 mazzy star和 dead can dance等。中文歌手以王菲为首，代表作有 “迷路”、“di-da”等。

(3.3)什么是 new age?

new age是种宁静、安逸、闲息的音乐，纯音乐作品占的比重较多，有歌唱的占较少。 new age可以是纯 acoustic(即以传统自发声乐器演奏)的，也可以是很电子化的，重点是营造出大自然平静的气氛或宇宙浩瀚的感觉，洗涤听者 的心灵，令人心平气和。 new age很多时与音乐治疗有关，不少new age音乐说可以治病，也有不少与打坐冥想有关，这与new age思潮哲学有莫大关系。 new age音乐通常被目为颇为中产阶级的音乐， windham hill是最具代表性的 new age唱片公司， new age代表乐手有 enya， george winston，william ackerman， yanni， kitaro等等。

(3.4)什么是electrophonic music?

何谓electrophonic music(电子音乐)?随着时代的演进，音乐家有了更多制作音乐的方法。所谓电子音乐，就是以电子合成器，音乐软体，电脑等所 产生的电子声响来制作音乐。电子音乐范围广泛，生活周遭常常能听到，在电影配乐，广告配乐，甚至某些国语流行歌中都有用到，不过 以电子舞曲为最。很多人认为电子乐是一种冷冰冰，没有感情的音乐。其实电子乐也可融入rock、 jazz甚至blues等多种元素而充满情感的。电子音乐的类型也是多种多样的，包括house 、techno、ambient、trance、psychedelic

trance、 breakbeat、 brit-hop、 big-beat、 trip-hop、drum’n'bass、 jungle、electro、dub、chill out、 minimalism。

(3.5)什么是techno ?

technology， 即表示“高科技舞曲”啦!利用电脑，合成器合成，做出一些特殊音效，这种音乐常常是许多音效组合起来的。 techno的节拍也是4/4拍,但速度较house快，且听起来具重复性，较强硬，较机械化，所以某些人称techno为“工 业噪音”，但某些还是会注重旋律的。 techno起源于美国底特律，有三名dj： derrick may，kevin saunderson juan atkins尝试将电子乐与黑人音乐结合，而产生了detroit techno。 detroit techno通常较平缓， 不像一般的techno那样强劲，可说是现代techno的起源。

(3.6)什么是ambient ?

听起来起伏不大，但其实一直在做改变，像是长时间的音效， 或是渐进式的音乐编排等等，常会营造出有层次的空间感，所以被称为“情境音乐”， 且常对於生活周遭的声音做取样，如人声、汽车声，甚至是其他音乐的旋律……等等。为70年代的brian eno所创，是一种很“高深”的电子音乐。后来ambient也有分支，如ambient house、 ambient techno……等等。相信大家从字面上就能了解，其实在很多种音乐中都会有ambient的影子，甚至某些古典乐中 也有ambient的味道。

(3.7)什么是trance?

迷幻舞曲，由techno演变而来，听了会让你有“出神”的感觉，但还是保有舞曲的律动，很注重bass的表现，某些听了会有“ 催眠”的效果。拍子也是以4/4拍为主。不过我也有听过用breakbeat的(下面会介绍)， 如sven vath的专辑fusion。

(3.8)什么是psychedelic trance?

又名goa trance，发源于印度的一个小岛上。goa即为这小岛的名字。旋律常带有印度风味。这类trance的特徵，就是很重视旋律 ，音符之间的细微变化，且常有重叠的旋律，和高亢的tb-303声音出现。有点ambient的味道。

(3.9)什么是break beat?

不像house的拍子boom-boom-boom-boom那么规律，而是以破碎的节拍呈现，像是切分拍。如在两拍中加入小碎 鼓，等等，breakbeat只是一个总称而已，有很多种音乐都属breakbeat，如以下的brit-hop、big-be at、和黑人的hip-hop便是一种。

(3.10)什么是brit-hop?

把黑人hip-hop的节奏加快鼓声加重，配合些许电吉他、bass，有摇滚乐的味道， 如the chemical brothers的曲风即是。

(4.1)什么是big-beat?

“大节拍”，把brit-hop变得更快，节奏更重，更有摇滚的味道，可说是将brit-hop改良后， 更成熟的一种曲风，你可以想像一下又重又碎，又有点快的节拍是什么样子。如the prodigy、the chemical brothers、 fat boy slim即是big beat的“大角”。

(4.2)什么是drum n’ bass?

对某些人来说，drum n’ bass的音乐可能很烦，因为drum n’ bass很注重在节奏上玩花样，还有bass的表现， 像是很快很快，复杂的碎拍，以及运用各种不同的电子鼓，或是取样而来的鼓声。因此drum n’ bass也是breakbeat的一种。4hero为drum n’ bass的知名乐团。

(4.3)什么是jungle?

jungle就是drum n’ bass的前身。jungle的速度比drum n’ bass快很多，而且节拍更为复杂，比较强悍，有时会搭配一些rap，不像drum n’ bass会融合一些jazz、soul等元素，听起来比较柔(当然这是和jungle比起来)。连“摇滚变色龙” david bowie现在都搞起jungle了。

(4.4)什么是electro?

早在70年代末，80年代初，那时还没有house、techno……这类玩意儿，电子音乐也较不盛行，只有electro这类 电子音乐而已。 electro可说是纯粹以电子合成器(synthesizer，长得有点像keyboard)来发声的音乐，通常蛮轻柔的，且 常使用roland公司的鼓机(dru

m machine)tr-808来发鼓声。德国的kraftwerk乐团可算是代表。 现在的electro音色比以前明亮，透明许多。后来一些dj利用electro变形成了一种hip-hop乐，被称作old school。现在的fatboy slim(流线胖小子)也融合了一些electro和old school在音乐中。

(4.5)什么是dub?

dub可说是将歌声抽离只剩下音乐的reggae(瑞格，也有称之为雷鬼，一种很有节拍性，唱腔特别的南美黑人音乐)。而将du b运用在电子乐上，则是将部份歌声抽离，将bass和鼓声加重，并且加入大量的echo(回音)等效果。这种技术常在现场表演时 ，将歌曲重新混音(remix)呈现。像massive attack 的专辑protection便由mad professor从trip-hop音乐混音成了dub版本。

(4.6)什么是chill out?

这只是一种统称而已。在舞池旁边，常会有一个区域，或是小房间，里头会放一些较缓和，较平静的音乐，让舞客跳完舞后缓和情绪，休 息一下，以放ambient音乐为主，所以会叫“chill out”(冷却)就是这个缘故。the orb可说是chill out音乐大师。

(4.7)什么是minimalism?

这被翻译成“极小派乐风”，在ambient和techno的音乐中常听的到。带有minimalism的曲子有着重复性的旋律 ，还有一种特别的律动，感觉上像一种特别的频率一样让你听了会有被催眠的感觉。jeff mills的作品就属于minimal detroit-techno，带有相当的迷幻色彩。

(5.1)black metal 黑色金属

以邪恶、异教崇拜以及撒旦思想为其主要诉求，通常在歌词含有反基督、反宗教的倾向即被归为黑金属，除了传统重金属固定乐器之外， 经常会使用其他乐器如钢琴、小提琴，或歌剧的女高音，在歌曲中营造出一股诡异又吓人的恐怖气氛。“黑金属”以来自北欧的重金属团 体最具代表性。

代表性乐队：emperor、 mayhem、 enslaved、 dark funeral

(5.2)death 死亡金属

乐风以“鞭击金属”或“碾核”为其背景演化而成。电吉他快速的反复，几无旋律的和弦，速击狂踩的双大鼓，主唱咬牙不清的低吟狂吼 ，歌词以死亡仇恨为主题，充满了尸体、内脏、肢解、分尸、奸尸、恋尸癖、食尸、虐待等变态字眼。“死亡金属”以来自佛罗里达州的 重金属乐队最具代表性。

代表性乐队：death、 carcass、 cannibal corpse、 suffocation

(5.3)doom 毁灭金属

节奏和速度都很慢，非常得慢。较不激烈也不大具有攻击性，但仍是相当拥有重量感的音乐。通常具有非常干净的音乐和歌声，但有时也 有粗暴野蛮的声音出现。

代表性乐队：solitude aeturnus、 cathedral、 candlemass

(5.4)glam metal 华丽金属

通常指一个重金属乐队的舞台表演方式或外在形象以夸张的浓妆艳抹或俊俏的外型来吸引乐迷，这类的乐队大多是偶像团体。“华丽金属 ”的乐风不会太重，歌词也无太深入之含义，属于较流行的重金属音乐，是主流金属的分支。

代表性乐团：kiss、 poison、 warrant、 twisted sister、 motley crue

(5.5)grindcore 碾核

类似“朋克”但极具重量的音乐，主唱低鸣深沉的狂吼咆哮、电吉他超麻辣的音色再加上极快速的反复节拍及密集鼓点，构成一种几无旋 律的重金属乐风，乐曲的长度都很短，从两秒钟到三分钟不等。“碾核”为一种极为冷僻的重金属乐风，大多数“死亡金属”乐队都采用 此一乐风。

代表性乐队：napalm death

(5.6)grunge 垃圾(颓废?)

另类摇滚的同义词。起源于西雅图的类金属(metal-like)另类摇滚，基本上grunge属于另类摇滚而不应归类到重金属 之下。

代表性乐队：nirvana、 pearl jam、 mudhoney、 alice in chains

(5.7)hardcore 硬核

源自于“朋克”，歌词直言不讳粗鄙不堪，音乐如同噪音一般。

(5.8)hard rock 重摇滚

比重金属乐“轻”一点的音乐。

代表性乐队：foreigner、 blue oyster cult、 heart、 cinderella

(5.9)heavy metal 重金属

“重金属”必须具备狂吼咆哮或高亢激昂的嗓音、电吉他大量失真的音色、再以密集快速的鼓点和低沉有力的贝司填满整个听觉的背景空 间，而构成一种含有高爆发力，快速度，重量感及破坏性等元素的改良式摇滚乐。通常泛指传统的主流派重金属或无法分类到其他重金属 流派里的重金属乐。

代表性乐队：ac/dc、 accept、 judas priest、 iron maiden、 w.a.s.p.

(5.10)industrial metal 工业金属

采用大量冰冷的电脑采样音效，穿插以机械或金属器具的撞击声代替传统打击乐器，加上电子式的节拍，是很象科技舞曲的重金属乐，但 仍保留重金属的大量失真效果。

代表性乐队：fear factory、 ministry、 nine inch nails

(6.1)neo-classical 新古典金属

受古典音乐极深，在重金属乐中加入大量古典音乐元素。

代表性乐队： angra

(6.2)pop metal 流行金属

在重金属摇滚中加入流行音乐之元素，可说是重金属摇滚和流行音乐妥协结合的折中产品，是极易为主流市场之音乐消费者所接受的重金 属摇滚乐。

代表性乐队：def leppard, mr big, white lion

(6.3)power metal 强力金属

拥有类似“速度金属”的速度和“鞭击金属”的重量压迫感，但旋律性不及“速度金属”，爆发力也不及“鞭击金属”。

代表性乐队：pantera, armored saint, wild dogs

(6.4)progressive metal 前卫金属

风格有别于传统重金属，在歌曲中采用大量复杂华丽的编曲，或在歌词中传达出某些前进思想供听者一个想象思考的空间。

代表性乐队：queensryche, fate warning, dream theater

(6.4)punk 朋克

歌词中传达某些叛逆思想及对生活环境、文化、社会、政治等的不满情绪，而音乐缺乏协调性无特定风格，是一种相当嘈杂的音乐，通常 一群能将乐器弄出声音来的人就可以组一个朋克乐队。

代表性乐队：clash, sex pistols

(6.5)speed metal 速度金属

速度为其标榜的主要特色。经常会和“鞭击金属”产生混淆，通常“速度金属”的音乐较具旋律性，主唱歌词咬字较清楚，吉他间奏富旋 律性且快速而流畅;而“鞭击金属”则较缺乏旋律性，完全一速度、重量、压迫感和破坏力并重。通常“鞭击金属”乐队和“速度金属” 乐队的乐手们都拥有纯熟高超的乐器演奏技巧。

代表性乐队：helloween, gamma ray, riot, rage

(6.7)thrash metal 鞭击金属

速度也极快，具有相当快速的反复节拍(riff)，电吉他粗麻的音色刷出剽悍的和弦，极具破坏力及压迫感的速踩双大鼓，主唱粗暴 狂吼式的唱腔配合电吉他快速的节奏急速地唱出几无旋律的曲调。“鞭击金属”以来自旧金山湾区的重金属乐队最具代表性。

代表性乐队：metallica, megadeth, slayer, anthrax, death angel, testament, exod

us, destruction, kreator, coroner, overkill

sunmoon 补充：

迪斯科(disco)

discotheque的简称，原意为唱片舞会，起先是指黑人在夜总会按录音跳舞的音乐，70年代实际上成了对任何时新的舞 蹈音乐的统称。与摇滚相比，它的特点是强劲的、不分轻重的、像节拍器一样作响的4/4拍子，歌词和曲调简单。1977年，因澳大 利亚流行音乐小组“比吉斯”(bee gees)的电影录音《周末狂热》在美国掀起迪斯科热。迪斯科经常在录音室进行音响合成，制成唱片，但终因节奏单调、风格雷同， 于80年代初逐渐被其他节奏不那么显著、速度稍慢的流行舞曲所代替。

朋克(punk)

70年代中期发展起来的一种摇滚乐，或称“朋克摇滚”(punk rock)，80年代初扩展到全美国。朋克针对摇滚乐的反抗性减弱这一情况，以一种故意与比较文明的摇滚乐相对立的面目出现。它 继承了摇滚乐开始时的反叛精神，蔑视一切传统，蔑视政府、宗教、教育机构等等，但却是用一种更加颓废的方式，极端无政府主义的言 行表现出来。他们的唱词淫狠、粗鲁;在舞台上尖叫、跺脚、斗殴;音乐上与硬摇滚比较接近，音响极大，音色浓烈浑厚。代表性乐队有 “性手枪”(sex pistols)、“撞击”(the clash)等。

70年代末，从朋克中分离出一支比较能被人接受的、也更带点艺术性的新品种，叫“新浪潮”(new wave)。

索尔(soul)

亦称“灵歌”。1969年，“公告牌”用“索尔”来代替原来对“节奏布鲁斯”的称呼。它是由布鲁斯、摇滚乐与黑人福音歌混合 而成的一种黑人流行音乐，演唱时较少演奏乐器。索尔的著名歌手有詹姆斯.布朗(james brown)、查尔斯(ray charles)、雷丁(ons redding)、史蒂夫.旺德(stevle wonder)、普林斯(prince)等。

雷鬼(reggae)

起源于牙买加，70年代中期传入美国。它把非洲、拉丁美洲节奏和类似非洲流行的那种呼应式的歌唱法，与强劲的、有推动力的摇 滚乐音响相结合。

说唱(rap)

黑人俚语，相当于“谈话”(talking)，产自纽约贫困黑人聚居区。它以在机械的节奏声的背景下，快速地诉说一连串押韵 的诗句为特征。这种形式来源之一是过去电台节目主持人在介绍唱片时所用的一种快速的、押韵的行话性的语言。莱普的歌词幽默、风趣 ，常带讽刺性，80年代尤其受到黑人欢迎。最有代表性的乐队是“公开的敌人”(public enemy)。莱普有时也称“希普-霍普”(hip-hop)。实际上，希普-霍普的含义更宽，泛指当时纽约街头文化的各种成分 ，除莱普外，还有：(莱普经常采用的)用手把放在唱

机转盘上的唱片前后移动，发出有节奏的刮擦声;唱片播放员(dj)在转换唱片 拼接唱片音乐片断时，听不出中断痕迹的技法;霹雳舞等等。

波普(pop)

一种具有强烈节拍的现代流行音乐。一般认为，流行音乐有广义和狭义之分，这里的“波普”是指狭义的流行音乐，如 70、80年代著名歌手罗斯(diana ross)、里奇(lionel richie)、惠特妮.休斯顿(whitney houston)、麦当娜(madonna)、迈克尔.杰克逊(michael jackson)、普林斯(prince)、珍妮特.杰克逊(janet jackson)等人演唱的很多歌曲。

艺术摇滚(art rock)

于70年代后流行，主要在欧洲，英国人领头，如“穆迪布鲁斯”(moody blues)乐队，但在美国也有影响。艺术摇滚把摇滚与古典音乐结合起来，利用电子设备演奏如李斯特、德彪西等作曲家的作品，又 称“进步摇滚”(prosressive rock)。

Seana Coulson

University of California, San Diego

Teenie Matlock

University of California, Santa Cruz

ABSTRACT

We propose an account of metaphor comprehension based on conceptual blending theory. We review data from on-line processing measures that support predictions of conceptual blending theory, and report results of an off-line feature listing study that assessed how different sorts of contexts alter the information activated by a given word. Participants generated features for words used in the null context, in sentences that promoted a literal reading of the target word, sentences that promoted a metaphorical reading, and sentences that required literal mapping. In literal mapping, the literal sense of the word was used in a way that prompts the reader to blend it with structure from a different domain. Results revealed some overlap in the features generated in each of the four contexts, but that some proportion of the features listed for words in literal, literal mapping, and metaphoric sentence contexts were unique and context-specific.

Metaphor and the Space Structuring Model

Characterizing the precise relationship between literal and nonliteral meaning dominates modern research on metaphor and figurative language. In traditional linguistic theory, literal and nonliteral meanings are seen as two different beasts, only one of which is well-behaved. On this view, “normal” language, that is to say, literal language, involves recruiting word meanings from the mental lexicon and combining them with grammar rules. Understanding normal language also demands compliance to communicative maxims: Utterances must be truthful, relevant, and maximally informative. In fact, on traditional accounts, conforming to these maxims is what enables speakers to discern literal language, which is thought to involve compositional parsing mechanisms, from nonliteral language, in which world knowledge and general reasoning processes must be invoked to understand the speaker’s intended meaning.

In the current work, we begin with a review of two influential approaches to metaphor processing, including the standard model of nonliteral language comprehension, and a competing model based on conceptual metaphor theory. In section 2, we offer our own account of metaphor comprehension based on the space structuring model (Coulson, 2001, in press), a theory of comprehension motivated by mental space theory (Fauconnier, 1994) and conceptual integration, or blending, theory (Fauconnier & Turner, 1998). In our model, metaphor comprehension involves coordinating various conceptual domains in a blend, a hybrid model that consists of structure from multiple input spaces, and that often develops emergent structure of its own. In sections 3 and 4, we review evidence consistent with our model, and discuss results of a feature listing study designed to assess some of its claims. In this study, people were asked to generate the features for a set of nouns used in a null context and in three types of sentence contexts that promoted a range of figurative readings. Quantitative analysis of these data shows that features produced in each sentential context differ from those for the same noun in the null context, and qualitative analysis reveals blending operations, such as elaboration. In section 5, we revisit the relationship between literal and nonliteral language in light of our results, and argue that both metaphorical and non-metaphorical meanings require the simultaneous activation of multiple cognitive models and the mappings among them.

1. Metaphor Processing

Classic literal and nonliteral distinctions are incorporated into the standard pragmatic model of metaphor processing (Grice, 1975; Searle, 1979), the validity of which has been a major focus of research on this topic (see Gibbs, 1994 for review). On the standard model, metaphor comprehension begins when the listener realizes that the speaker has intentionally violated the Gricean Maxim of Quality, “Be truthful.” Upon realizing the literal incongruity of a metaphoric utterance, the listener must then derive a nonliteral interpretation. Consequently, the standard model suggests that understanding metaphoric language takes longer than non-metaphoric language, and involves qualitatively different processes (see Gibbs, 1994, Gibbs & Matlock, in press).

1.1 The Standard Model and Conceptual Metaphor Theory

Cognitive linguists have challenged many of the traditional assumptions about literal and nonliteral language. In particular, conceptual metaphor theory (CMT) proponents have shown that metaphor is not merely a literary device, but an integral part of everyday language and thought (Lakoff & Johnson, 1980; Sweetser, 1990; Turner, 1991). Based on linguistic patterns that turn up in language after language, CMT suggests a principled relationship between literal and nonliteral language, with evidence that metaphoric meanings are systematically related to literal ones. For instance, countless clusters of expressions use the same kinds of words to talk about very different experiential domains. This is seen in the way people describe love in terms of travel, as with cruise and crash in, “Their relationship was cruising along but suddenly crashed,” (see Gibbs, 1994; Gibbs & Nascimento, 1993), or in the way they use words referring to vision to express understanding, as with see in “I see what you’re saying,” (see Sweetser, 1990). This tendency is also apparent in the way verbal arguments are described in terms of physical battles, as in “He attacked every weak point in the argument,” (Lakoff & Johnson, 1980).

To explain this systematicity, Lakoff and Johnson (1980) propose that metaphors reflect the output of a cognitive process by which we understand one domain, known as the target, by exploiting cognitive models from an analogically related domain known as the source. The systematicity in the use of source and target domain terminology derives from the fact that some of the logic of the source domain has been imported into the target in a way that maintains the mappings from one to the other. Thus construed, metaphoric language is the manifestation of conceptual structure organized by a cross-domain mapping: a systematic set of correspondences between the source and target that result when cognitive models from a particular source domain are used to conceptualize a given target domain.

These observations point to the inadequacy of the standard model as a comprehensive account of metaphor comprehension. The model fails to explain both the ubiquity of systematic correspondences and the logic of metaphorical expressions. It also fails to explain current metaphorical use and the development of well-documented cross-linguistic patterns. While CMT proponents appeal to shared cognitive models to explain metaphor comprehension, current vocabulary use, and semantic change (e.g., Lakoff, 1993), the standard model leaves these details unexplained.

1.2 Experimental Evidence

Besides linguistic evidence against the literal/figurative dichotomy, the standard pragmatic model is also undermined by a good deal of experimental evidence. First, empirical work refutes the assumption that literal processing is obligatory and necessarily prior to metaphoric processing (e.g., Glucksberg, Gildea, & Bookin, 1982; Keysar, 1989). Psycholinguists have also challenged the prediction that metaphoric meanings take longer to compute than literal ones by contrasting reading times for both types of statements. While reading times for metaphors are generally longer in minimal contexts, when the same stimuli are embedded in longer passages that provide supporting context, literal and metaphorical utterances are read and understood at the same rate (Inhoff, Lima, & Carrol, 1984; Ortony, Schallert, Reynolds, & Antos, 1978).

Such results go against the standard model, but are well-explained by one fairly controversial model of metaphor processing, the direct access model (Gibbs, 1994). Motivated in part by conceptual metaphor theory, this model holds that metaphor comprehension requires the same processes as the comprehension of literal language. The direct access model holds that difficulty in processing metaphoric language is a function of contextual support for the recruitment of the cross-domain mapping or mappings needed to understand any given metaphor. On this view, while literal meanings may tend to predominate in the interpretation of decontextualized utterances, metaphoric meanings require realistic social contexts. Controversially, the direct access model maintains that context can even bias a metaphoric meaning over a literal one.

Although the direct access model finds support in the finding that the nonliteral meaning of familiar idioms is almost immediately available, it is undermined by various reports that literal aspects of word meaning are primed even in metaphorical contexts. For example, using a word fragment completion task, Giora & Fein (1999) found that both literal and metaphoric meanings were activated in the comprehension of familiar metaphors. Similarly, using the cross-modal priming technique, Blasko and Connine (1993) found priming for the literal as well as the metaphoric meanings in familiar metaphors. For unfamiliar metaphors, they found priming only for the literal meanings of their stimuli. Moreover, in the processing of unpredictable idioms, Cacciari and Tabossi (1998) report priming for literal meanings immediately at the offset, and for both literal and nonliteral meanings 300 ms later.

2. Conceptual Integration and Metaphor Comprehension

Our own model of metaphor comprehension, the space structuring model (SSM), also acknowledges the prevalence of metaphor in everyday language and thought, as well as commonalities between the conceptual basis of poetic language and the conventional metaphors described by cognitive linguists (e.g., Lakoff & Turner, 1989; Turner, 1996). Like many models of metaphor comprehension, SSM also advocates commonalities in the construction of literal and nonliteral meanings. However, besides conceptual metaphor theory, SSM is directly motivated by conceptual blending theory (Coulson, 2001, in press; Fauconnier & Turner, 1998). Blending is a set of operations for combining cognitive models in a network of mental spaces (Fauconnier, 1994; Fauconnier & Turner, 1998). In SSM, comprehension involves the temporary construction of simple cognitive models along with the establishment of mappings, or, systematic correspondences among objects and relationships represented in various models. Mappings are based on pragmatic functions such as identity, similarity, or analogy. Consequently, metaphoric meanings – that use analogy to link objects in different spaces — do not fundamentally differ from meanings that employ other sorts of mappings.

2.1 Mental Spaces

In SSM, linguistic cues prompt speakers to set up elements in mental spaces, a level of referential structure whose contents need not refer to objects in the world (Fauconnier, 1994). A mental space can be thought of as a temporary container for relevant information about a particular scenario as perceived, imagined, remembered, or otherwise understood. Initially devised to address indirect reference and referential opacity, mental space theory has proven to be useful for semantic and pragmatic complexities (see Fauconnier, 1997; Fauconnier & Sweetser, 1996). For instance, mental spaces can represent examples in which Titanic refers to both the ship and the movie about the ship, as in, “ Titanic is a movie about the voyage of the Titanic.” By partitioning the information in this sentence into two linked spaces, mental space theory captures the fact that though the ship and the movie differ, the correspondence between them is not completely arbitrary.

Mental space theory was initially designed to keep incompatible information about a single object in discrete representations, for instance, a girl with green eyes in reality could have blue eyes in a picture. But the more recent theory of conceptual integration posits a particular kind of mental space, a blended space, in which this sort of incompatible information is brought together to generate inferences that can be projected to other spaces (Fauconnier & Turner, 1998). For example, blended spaces can represent expressions using structure from multiple spaces, as with the headline “Titanic: Unsinkable after all.” In contrast to the previous example, in which the film and the ship are clearly distinguished, the headline exemplifies simultaneous reference to the ship, claimed by some to be unsinkable, but which proved otherwise, and the movie about the ship, which proved to be quite successful, both with the critics and the general populace.

2.2 Conceptual Integration Networks

A computational- (though not algorithmic-) level account of blending appeals to a conceptual integration network, an array of mental spaces (Fauconnier & Turner, 1998). Blends have two or more input spaces structured by information from discrete cognitive domains, a generic space that contains abstract structure common to all spaces in the network, and a blended space that contains selected aspects of structure from both input spaces, as well as emergent structure of its own. For example, in the unsinkable Titanic blend, one input space contains information about the historic ship (which sunk, and therefore was not unsinkable), while the other input contains information about the movie (which did well). Though one does not usually talk about whether movies are good flotation devices, the conceptual structure in these input spaces can nonetheless be aligned via analogical mappings between the ship and the movie, the ship’s voyage and the movie’s run, and between the ship’s fate (sinking) and the movie’s fate (winning oscars). Blending theory differs from CMT in that it explicitly allows for disanalogies in the representation of metaphoric expressions.

Elements in each of the four spaces in the integration network for the Titanic blend are shown in Table 1. The generic space in this network contains a schematic representation of the common event structure, that is, an unspecified agentive object that undertakes a course with an unspecified purpose, and whose outcome can be successful or unsuccessful. Conceptual structure in the two input spaces, then, are analogically linked, while the mappings between the inputs and the generic space involve category inclusion. The blended space, too, shares the abstract event structure in the generic space, and is composed of a combination of some structure from each of the input spaces. In this example, the blended space inherits some structure from the scenario associated with the historic input, and some structure from the movie input, in particular, the fate of the movie. The mappings between the ship and the voyage in the blended space and the ship and the voyage in the historic space are identity mappings. However, the successful voyage of the Titanic in the blended space maps onto the success of the movie via analogy mappings (see Turner & Fauconnier, 2000 for more Titanic blends). Integrating a representation of the Titanic’s voyage with the fate of the movie yields a counterfactual rendering of the Titanic’s voyage in which the ship does not sink.

2.3 Conceptual Blending and Metaphor Comprehension

Following Fauconnier and Turner’s conceptual integration theory (1998), we argue that metaphor is more than a set of mappings between a source domain and a target domain. On our view, metaphor involves a complex of mappings with multiple spaces in conceptual integration networks. SSM differs from a number of other models of metaphor comprehension in that it does not posit the existence of a discrete metaphorical meaning. Rather, metaphorical meaning arises out of the information represented in the integration network. For instance, understanding the metaphor in “All the nurses at the hospital say that surgeon is a butcher,” requires coordinating conceptual structure associated with surgery, butchery, and a blend of the two (Grady, Oakley, & Coulson, 1999).

As in CMT, comprehension of the butcher metaphor requires one to apprehend the mappings between surgeon and butcher, patient and dead animal (e.g. cow), as well as scalpel and cleaver. However, it also involves construction of a blended space in which structure from each of these inputs can be integrated. In this example, the blended space inherits goals of the surgeon, and the means and manner of the butcher. The inference that the surgeon is incompetent arises when these structures are integrated to create a hypothetical agent with both characteristics. Behavior that is perfectly appropriate for a butcher whose goal is to cut up a dead cow is indeed appalling for the surgeon operating on a live human being.

Integration in the blended space involves three related processes, composition, completion, and elaboration, each of which provides for the possibility of emergent structure. Composition involves attributing a relation from one space to an element or elements from the other input spaces. Composition can be as simple as integrating an element (such as dinner) with a frame (such as four-course), or can involve more creative blending, as in the integration of frames for Irish, and four-course, with dinner (three pints of Guinness and a bag of crisps). In either case, emergent structure arises from the contextual accommodation of a concept from one domain to apply to elements in a different domain. Completion is pattern completion which occurs when structure in the blend matches information in long-term memory. For instance, if a friend told you that he had gone to Baskin Robbins for ice cream, you might infer that he had eaten a cone there as well. Elaboration, related to completion, involves mental simulation of the event represented in the blend. For example, we suggest that the following excerpt from a performance report is funny because the reader mentally imagines the scene, “Since my last report, this employee has reached rock bottom and has started to dig.”

We suggest that speakers exploit explicit grammatical cues to construct a blended space with conceptual structure from both input domains. Metaphor comprehension thus involves the activation of conceptual structure needed to construct the model in the blended space, the activation of conceptual structure in the input and generic spaces, and the establishment of mappings between spaces in the network. Emergent structure is activated in order to produce a relatively coherent juxtaposition of disparate aspects of conceptual structure from the input domains. Moreover, particular inferences that issue from the use of a given metaphoric expression reflect the fact that metaphoric projections recruit processes of conceptual blending to produce emergent structure that can be mapped back onto the inputs.

3. Processing Metaphoric Language

The SSM makes a number of predictions for on-line meaning construction. For instance, because it is based on a general theory of conceptual integration, SSM suggests the same conceptual operations are involved in the comprehension of literal and nonliteral language. For example, understanding butcher in “During the war, that surgeon had to work as a butcher,” requires the comprehender to set up simple cognitive models in mental spaces, and establish mappings based on shared relational structure. As in metaphoric uses of butcher discussed in the previous section, inferences are generated in the blended space, where information about a surgeon’s training and skill is integrated with general information about butchers, or other aspects of the context. One might, for instance, infer that the surgeon in question was overqualified for his job, or that he was forced to work as a butcher in a labor camp.

Like many modern models of metaphor processing (see Giora, 1997 for review), the SSM suggests that qualitatively similar processing operations underlie the comprehension of literal and nonliteral meanings. Consequently, the model is supported by evidence that metaphoric meanings are understood in approximately the same amount of time as literal control statements. Moreover, findings from a small set of on-line studies demonstrate that variables pertaining to difficulty of processing metaphoric items also pertain to the difficulty of processing literal items. For instance, familiarity, one such variable, is a determinant of processing difficulty for literal and non-literal language alike (Gernsbacher, 1984). In a cross-modal priming study, Blasko and Connine (1993) show that the familiarity of a metaphor affected reaction times for words related to its metaphorical meaning. In an eye-tracking study, Blasko and Briihl (1997) found that gaze durations for metaphorical expressions decreased both as a function of familiarity and as a function of contextual support. Similarly, Frisson and Pickering (1999) found equivalent gaze durations for sentences containing interpretable metonymies and sentences containing literal interpretations of the same words (see also Frisson & Pickering, this volume).

McElree and Nordlie (1999), however, argue that the presence or absence of differences in reading times can result from a number of different factors, not all of which reflect true differences in processing time. One way to tease apart stimulus-related processing from decision-related processing is to measure the speed accuracy tradeoff (SAT) curves as participants perform a judgment task at varying amounts of processing time. With adequate sampling, it is possible to observe the full timecourse of processing by establishing the point in time when performance exceeds that of chance, the point at which performance reaches an asymptotic level, and the slope of the curve between the former and the latter. Using SAT to investigate the timecourse of meaning activation in literal and metaphorical statements, McElree and Nordlie (1999) found no evidence of literal meanings being available earlier than figurative meanings.

Moreover, event-related brain potential (ERP) data support the claim in the direct access model that difficulty in the comprehension of metaphoric utterances is largely a function of contextual support (Pynte, Besson, Robichon, & Poli, 1996). This latter finding is especially important because the ERP methodology can address some limitations of chronometric studies. As Gibbs (1994) notes, parity in reading times for literal and metaphorical expressions need not entail parity in the underlying comprehension processes. It is possible, for example, that literal and metaphorical meaning might take the same amount of time to comprehend, but that the latter required more effort or processing resources (Coulson & Van Petten, 2000 submitted). Alternatively, comprehension processes for literal versus metaphoric utterances might take the same amount of time to complete, and yet involve quite different computations (Gibbs & Gerrig, 1989).

3.1 ERPs

Because they involve a direct and continuous measure of brain activity, event-related potentials (ERPs) can potentially distinguish between qualitatively different sorts of processing, even if their corresponding behavioral manifestations require the same amount of time (see Coulson, King, & Kutas, 1998 for review). ERPs are small voltage fluctuations in the EEG that are timelocked to sensory, motor, or cognitive events collected by recording EEG while participants perform a cognitive task such as reading (Rugg & Coles, 1995). By averaging the EEG timelocked to multiple tokens of a given type (e.g. the onset of a word used metaphorically), it is possible to isolate aspects of the electrical signal that are temporally associated with the processing of that type of event (such as understanding a metaphoric meaning). The result of averaging is a waveform with a series of positive and negative peaks, known as components labeled by reference to their polarity (’P’ for positive-going and ‘N’ for negative-going), and when they occur relative to the onset of the stimulus event, or relative to other ERP components.

One ERP component of particular interest to researchers interested in meaning is the N400, so-called because it is a negative-going wave that peaks approximately 400 ms after the presentation of a meaningful stimulus. The N400 was first noted in experiments contrasting sentences that ended sensibly and predictably with others that ended with an incongruous word. Congruous words elicited a late positive wave, while incongruous endings elicited a negative wave beginning about 200 ms after the stimulus was presented and peaking at 400 ms post-stimulus (Kutas & Hillyard, 1980). Subsequent research indicates that N400 is elicited by all words, written, spoken, or signed, and that N400 amplitude indexes the difficulty of integrating a word into the established context (see Kutas, Federmeier, Coulson, King, & Muente, 2000 for review). The greater the processing difficulty associated with a word, the larger the N400 component it elicits (see Figure 1, and note that negativity is plotted up).

Taking advantage of this well-known interpretive feature of the N400, Pynte and colleagues contrasted ERPs to familiar and unfamiliar metaphors in relevant versus irrelevant contexts. They found that regardless of the familiarity of the metaphors, N400 amplitude was a function of the relevance of the context. Moreover, by using ERPs Pynte and colleagues employed a measure which is in principle capable of revealing the qualitative processing differences predicted by the standard model. In fact, they observed no evidence of a qualitative difference in brain activity associated with the comprehension of literal and metaphoric language.

Reports that literal and nonliteral language comprehension both display a similar timecourse and recruit a similar set of neural generators are consistent with predictions of the SSM. Moreover, the SSM also makes predictions for comprehension difficulty, predicting a gradient of processing difficulty related to the extent to which the integration requires the comprehender to elaborate the scenario set up in the blended space. This prediction was tested by Coulson and Van Petten (2000, submitted) when they compared ERPs elicited by words in three different contexts on a continuum from literal to figurative, as suggested by blending theory. For the literal end of the continuum, Coulson and Van Petten used sentences that promoted a literal reading of the last term, as in, “He knows that whiskey is a strong intoxicant.” At the metaphoric end of the continuum, they used sentences which promoted a metaphoric reading of the last term, as in, “He knows that power is a strong intoxicant.” Coulson and Van Petten also posited a literal mapping condition, hypothesized to fall somewhere between the literal and the metaphoric uses, such as, “He has used cough syrup as an intoxicant.”

Literal mapping stimuli employed fully literal uses of words in ways that were hypothesized to include some of the same conceptual operations as in metaphor comprehension. These sentences described cases where one object was substituted for another, one object was mistaken for another, or one object was used to represent another – all contexts that require the comprehender to set up mappings between the two objects in question, and the domains in which they typically occur. In line with many models of metaphor comprehension (e.g., Gibbs, 1994; Giora, 1997; Glucksberg, 1998), the space structuring model predicts qualitatively similar brain responses to literally and metaphorically used words, suggesting the same processes are used in literal and nonliteral language comprehension. Further, in positing a continuum from literal to metaphorical based on the difficulty of the conceptual integration needed to comprehend the statement, blending theory predicts a graded difference in N400 amplitude for the three sorts of stimuli.

Overall, data reported by Coulson and Van Petten (2000, submitted) were largely consistent with the predictions of the SSM. In the early time window, 300-500 ms post-onset and before, ERPs in all three conditions were qualitatively similar, displaying similar waveshape and scalp topography. This suggests that during the initial stages, processing was similar for all three sorts of contexts. Moreover, as predicted, N400 amplitude differed as a function of metaphoricity, with literals eliciting the least N400, literal mappings the next-most, and metaphors eliciting the most N400, suggesting a concomitant gradient of processing difficulty. The graded N400 difference argues against the literal/figurative dichotomy inherent in the standard model, and suggests processing difficulty associated with figurative language is related to the complexity of mapping and conceptual integration.

4. Feature Study

In their ERP study, Coulson and Van Petten (2000, submitted) show a processing gradient, which they attribute to the complexity of blending operations needed to understand words in the literal, literal mapping, and metaphorical contexts. However, aside from the authors’ native speaker intuitions, there was no evidence to show that placing these words in different sentential contexts would promote the retrieval of different sorts of conceptual structure, as hypothesized in the space structuring model. Indeed, a general characteristic of research that addresses the issue of continuity between processes underlying literal and metaphoric language comprehension is that it fails to address the details of metaphor comprehension. However, another way of addressing the relationship between both sorts of meaning construction is to examine the information that people activate when they understand literal versus nonliteral language.

This is the approach taken by Tourangeau and Rips (1991), in a study that compared the sorts of features people listed for metaphoric language with those listed for the contributing source and target domain concepts. Tourangeau and Rips found that many of the features listed for the metaphoric meanings were emergent, that is, they were not established parts of either of the domains in the metaphor. For instance, respected was listed as a feature of the eagle in “The eagle is a lion among birds,” but was not listed as characterizing either eagles or lions when considered independently (Tourangeau & Rips, 1991). Further, their participants rated the emergent features as being more crucial to the meaning of the metaphor than, for example, features that people listed for both eagles and lions. Tourangeau and Rips (1991) suggest that this pattern of data argues against models, such as Gentner’s (1983; Gentner & Wolff, 1997, Wolff & Gentner, 2000) structure mapping engine and Glucksberg and Keysar’s (1990) property attribution model, that posit the computation of shared features as the basis of metaphor comprehension.

Like Tourangeau and Rips (1991), we suggest that metaphor comprehension requires the transformation rather than pure transfer of properties from one domain to another. Moreover, the transformation occurs via blending processes such as completion and elaboration. In positing continuity between literal and nonliteral meaning construction, the SSM predicts that “emergent” features should arise in the course of conceptual integration across the continuum from literal to figurative meanings. Consequently, we conducted an off-line study that compared the sorts of features participants generated to words in a null context with the features they listed for the same words in literal, literal mapping, and metaphoric contexts of the sort employed by Coulson and Van Petten (2000, submitted).

In the current study, we are primarily concerned with the role of sentential context in the construction of meaning, especially how manipulating the context in which a word appears can influence the interpretation of that word, as determined by the features participants produce. One possibility is that participants would generate the same features for a word, regardless of the context in which it appeared. Such a result would suggest the construction of word meaning is removed from contextual integration, being identical from context to context. Alternatively, people might generate features relevant to and reflective of the particular sentential context in which they occur. This pattern of responses would indicate that people integrate contextual factors in such a way as to alter their understanding of individual words. Further, in a qualitative analysis of features participants generate, we should expect to see evidence of blending processes such as completion and elaboration in all three sorts of contexts.

Method

Design, Stimuli, and Participants. The study was a within-subjects design with four conditions, including a null context and three sentential contexts. In the null context, the target word appeared in isolation. In the sentential contexts, the target word appeared at the end of a sentence context. In the literal condition, the target word appeared in its literal sense, as with anchor in, “Last time he went sailing he almost forgot about the anchor.” In the metaphoric context, the target word appeared in its metaphorical sense, as with anchor in, “Amidst all the trappings of success, his wife was his anchor.” The literal mapping condition is served as an in-between condition, whereby the target word was used in its literal sense, but appeared in a context requiring the reader to perform some of the same integration operations hypothesized to underlie metaphor comprehension. For example, the literal mapping stimulus for anchor was, “We were able to use a barbell for an anchor,” in which a barbell has been projected into the sailing scenario to fulfill the function of an anchor.

The 35 words in this study were embedded in a larger feature listing study which included 12 lists seen by 120 UCSC undergraduates, all fluent English speakers. In the null context, each word was seen and rated by 20 participants. In the sentential contexts, each word was seen by 10 participants in each of the three types of sentences. Stimuli were distributed across lists such that no participant saw the same item in more than one context.

Procedure. Participants were given a booklet with two sections: Part A, a list of words (null context condition), and Part B, a list of sentences (randomly ordered items from three sentential context conditions). In Part A, participants read each item and jotted down two to three features or characteristics of that item. In Part B, they read each sentence and quickly listed two or three features for the underlined word. Participants were told that they were not being timed, but encouraged not to dwell on any one item. When unsure about the meaning of a word, they were to leave a blank.

Results. For each of the 35 stimuli, participants’ responses were compiled into a file that contained a list of features generated for that word in the null context, and in each of the three sentence conditions. Data were quantified in two ways, one a measure of the proportion of unique features in each condition, and one a measure of the similarity of the features for words in different sentential contexts. First, for each of the three sentence types, we calculated the proportion of features that were unique to that condition, viz. not produced for any of the other conditions. When words were presented in literal contexts, 41.77% of the features were not generated in either of the other sentential contexts or for the same words in the null context. When words were presented in literal mapping contexts, 39.66% of the features were unique to that context. Finally, when words were presented in the metaphorical context, 46% of the features were unique.

As is evident in Figure 2, metaphors elicited reliably more unique features than the other two (literal) sentence types. Nonetheless, placement of the stimuli in all three sorts of sentences resulted in the elicitation of a substantial proportion of unique features. The high proportion of unique features in each of the sentence contexts (ranging from 40% to 46%) suggests a remarkable degree of context-sensitivity in the conceptual structure participants retrieved for these materials. Although the off-line feature listing task cannot assess whether participants actually use this information during comprehension, the generation of unique features indicates a systematic difference across conditions in the availability of the information that the participants considered relevant. These differences suggest a word’s appearance in any sentential context can modulate which aspects of conceptual structure participants are likely to exploit in meaning construction. This was especially the case for sentential contexts that promoted a term’s metaphorical meaning.

However, it is potentially misleading to focus on the percentage of unique features. For example, it is possible that participants listed different words to express characteristics of the stimuli in each of the sentential contexts, but that the conceptual differences denoted by those words were minimal. For this reason, we assessed the similarity of the feature sets elicited by stimuli in each sentence type by using the latent semantic analysis (LSA) method, a method for creating statistical profiles of linguistic items, via the representation of words in a high dimensional semantic space derived from statistical analysis of large text corpora (see Landauer, Foltz, & Laham, 1998). By extracting multivariate correlation contingencies between a word and its context, LSA produces representations whose relative proximity in semantic space can be shown to closely mimic human judgments of semantic similarity (Landauer & Dumais, 1997).

To assess the semantic similarity of the feature sets elicited in our study, we transformed each feature set into a vector in a high dimensional semantic space (300 dimensions) derived from latent semantic analysis (LSA) of a large corpus (119,627 paragraphs) of machine readable texts, including novels, newspaper articles, and educational texts. This yielded four vectors for each word, one that represented the null context features, and one for each of the literal, literal mapping, and metaphorical feature sets. Semantic similarity was assessed by measuring the cosine of the angle between the vectors in each sentence condition to the vector representing the null context feature set. The cosine thus functions as a measure of proximity in semantic space, where 1 is identity, and 0 represents orthogonal vectors.

The average similarity score was 0.84 between the null context and the literal feature sets, 0.81 for the null context and the literal mapping feature sets, and 0.78 for the null context and the metaphorical feature sets. These scores indicate that the features listed for words in the metaphorical contexts were the least similar to those listed in the null context, words in the literal contexts were the most similar, and words in the literal mapping contexts fell somewhere in between. Repeated measures analysis of variance on cosine measures revealed a main effect of sentence context, F(2,68)=5.48, p<0.01, but post-hoc comparisons suggested that while the literal and metaphorical measures differed reliably from each other, t(1,34)=3.26, p<0.01, the literal mappings did not differ from either the literal, or the metaphorical measures. This result is consistent with the assumption that the literal mapping stimuli were intermediate with respect to literal and metaphorical stimuli.

Our analysis also included examination of unique features generated for a few words in the three sorts of sentence conditions. Although participants were specifically instructed to focus on the word at the end of the sentence, many features listed were apparently influenced by previous context. For example, with “Unfortunately, what started as a mere flirtation with the stock market has become an orgy,” participants generated unique responses such as EXCESSIVE, CROWDED, INDULGENT, that might be classified as low-salient properties of orgies. However, they also listed CONFUSING, COMPLICATED, and EXPENSIVE. These negative properties are clearly influenced by context, such as the word unfortunately, and the integration of concepts related to orgies with concepts related to the stock market.

Moreover, evidence of integration was not limited to contexts that promoted a metaphorical reading. It was also observed in the literal mapping and literal contexts for orgy. For the literal mapping context, “He saw some hippies headed for the river and assumed it was an orgy,” participants listed unique features such as 70’s, DRUGGIES, SMOKING, WOODS, and SKINNY DIP, which clearly reflect concepts related to context, including hippies and river. It is reasonable to assume that such responses reflect the process of elaboration, or imaginative simulation of what the hippies might do or how they might behave. Similarly, features generated in the literal context, “They ended the year with a huge party that everyone remembered as the orgy,” also show the influence of context. For instance, unique responses for orgy in the literal context include FOOD and DRINK, items not normally associated with the canonical meaning of orgy, but which emerge through completion of the party scenario.

In the metaphor “The coach said he’d miss his seniors because they were the backbone,” responses included RELIABLE, SECURE, and RIGID, as well as BEST and FASTEST, which were clearly influenced by integration of information about the role of backbones in vertebrates and the role of the seniors on the coach’s team. Examples such as this underline the importance of the relational structure shared between the input domains, as emphasized in Gentner and colleagues’ model of metaphor comprehension (Gentner & Wolff, 1997, Wolff & Gentner, 2000). While the SSM also maintains an important role for analogical mapping in metaphor comprehension, the presence of shared relational structure is not as essential for our model as for Gentner and colleagues. In fact, the SSM predicts that people can comprehend metaphorical meanings that involve explicit disanalogies between the input domains.

Responses for the literal mapping context, “The paleontologists quickly discovered that the foot bones were actually fragments of backbone,” included BREAK, BROKEN ARMS, DELICATE, and INJURY, features that have little or nothing to do with backbones per se. Once again it is apparent that context influenced the features participants produced. We suggest that fragments drove the choice of responses in these examples, and that the people who listed these features used elaboration to produce a scenario to explain why the bones were fragmented. Responses for the literal context, “At the academy, young FBI officers are taught to target the backbone,” include VULNERABLE, IRREPARABLE, and DAMAGING, which involves the integration of information about FBI officers with what it means to target a backbone, and completion of the integrated scenario. Other examples of features generated are shown in Table 2.

In sum, we found that there are differences but also similarities in the types of features generated in each context. In particular, metaphorical sentences elicited more unique features than the other two conditions, but the overall high proportion of unique features generated in all sentential contexts suggests a good deal of context-sensitivity. At the same time, though, we have to acknowledge that the similarity across the feature sets was quite high. Approximately 60% of the features listed in each sentential context were also listed in the null context, indicating some degree of constancy in the conceptual structure available for meaning construction. Therefore, we can assume that when a word appears in a sentential context, the presence of the word and its interaction with the context can alter or drive certain aspects of conceptual structure, which are exploited in meaning construction. We attribute the systematic differences in the types of features produced in various sentential contexts to differences in blending operations. In particular, as noted, literal and literal mapping stimuli tended to engender completion, while metaphorical stimuli were more likely to engender elaboration.

5. As time goes by

In positing continuity between literal and nonliteral meaning construction, the SSM is supported by the consistent finding that when contextual factors have been equated, literal and metaphoric meanings take the same amount of time to compute. The SSM is also supported by research that indicates that variables such as familiarity and contextual support influence the processing difficulty of both literal and nonliteral language. Further, ERP data suggest that the same set of brain regions mediate the construction of both literal and nonliteral meanings. However, continuity between literal and nonliteral language processing is a feature of most modern models of metaphor comprehension. Consequently, evidence that supports the SSM also supports the direct access model in which metaphorical meanings can be activated independently of literal ones (Gibbs, 1994), a parallel model in which neither the literal nor metaphorical interpretation has priority (Cacciari & Glucksberg, 1994; Glucksberg, 1991), and an underspecification model in which the processor initially activates the same underspecified representation for literal and figurative meanings, and only later does it fill in the details (Frisson & Pickering, 1999, and this volume).

However, the SSM finds more support in the ERP data reported by Coulson and Van Petten (2000, submitted). Although the direct access model is supported by the similar timecourse of ERPs elicited by metaphoric and literal uses of the same words, it is undermined by quantitative differences in the N400 that indicate metaphors are harder to process. This finding also argues against the underspecification model (Frisson & Pickering, 1999).[1] If the parser employs a single underspecified representation each time it encounters a word, processing difficulty should be independent of figurativity, and thus predicts equivalence in N400 amplitude as well as in gaze durations. While the gradient of processing difficulty, from literal to literal mapping, to metaphorical, might be consistent with other models of metaphor comprehension, it is most directly implied by the theories of blending and mental spaces.

Interestingly, the processing difficulty gradient observed by Coulson and Van Petten (2000, submitted) was paralleled to a certain extent by the similarity gradient of the different feature sets participants generated in the current study. Comparing features that people generated for words used in contexts that promote the same range of figurative meanings as in Coulson and Van Petten (2000, submitted), we found that literal meanings were most similar to the information associated with a word in the null context, literal mappings the next-most, and metaphorical meanings were the least similar. This presents the possibility that the observed difficulty gradient relates to blending operations needed to activate the features that were unique to each context. Qualitative analysis of these unique features indeed suggests that while there is evidence for all of the blending processes in each of the conditions, literal uses tend to engender composition and completion, while metaphorical uses were more likely to promote elaboration.

Of current models of metaphor comprehension, the SSM is most similar to the model proposed by Cacciari and Glucksberg (1994), especially in being a parallel model. Perhaps it is not surprising, then, that the SSM is supported by findings in a study by Cacciari and Glucksberg (1995) in which participants were asked to describe mental images formed in conjunction with the comprehension of a number of Italian idioms normed for familiarity and for opacity, or the extent to which its literal and figurative readings were related to one another. Interestingly, Cacciari and Glucksberg report descriptions of imagery judged as figurative which seem to us to represent the sorts of images associated with a blended space. For example, for the Italian idiom lose one’s head, which means to become crazy, so-called figurative depictions of this idiom included, “I am laughing to tears and I lose my head in a jump, the head jumps away,” as well as “A crazy person that no longer has control over his actions, his head is empty, transparent, without its content, the brain,” (Cacciari & Glucksberg, 1995: 50-51).

Both these examples are characteristic of cognitive models represented in the blended space in a conceptual integration network. While not all blends are chimerical, it is not unusual for unrealistic, impossible events such as a headless body, or a head without a brain, to be represented in the blended space. The first example, in which we have a head that spontaneously separates itself from its body, can be represented in a conceptual integration network in which one input contains a model of the realistic implications of a head falling off (death), and the other contains a model of an unspecified cause resulting in erratic behavior. The blended space inherits the cause from the first input, and the effect from the second, such that the head falling off the body is understood to cause erratic behavior. In the SSM, the meaning of metaphoric language is not represented in any single space in the integration network, nor in any single analogical link, but emerges from apprehension of the relationships among the various elements in the network.

Cacciari and Glucksberg (1994) argue that evidence for the activation of literal meanings in metaphorical context, reflects the parallel activation of both sorts of meanings. In contrast, the SSM explains such data by pointing to the principled relationship between literal and nonliteral meaning in conceptual blending theory. While researchers in CMT argued that the literal content of metaphorical expressions indicates congruity between both the language and the logic of the source and target domains, conceptual blending theory takes this observation one step farther in arguing that the mixture of source and target domain language in metaphoric utterances is mirrored in the logic of the blend. Indeed, blending theory is in part motivated by the observation that speakers often employ source domain language without fully utilizing source domain logic.

Evidence for the import of blending in metaphoric langauge can be found in examples that contain partial disanalogy (e.g. Coulson, 1996). For example, the presence of disanalogy is particularly common in idioms like digging your own grave, (discussed in Coulson, 1997; Fauconnier & Turner, 1998). Coulson (1997) shows how various instances of the metaphoric idiom digging your own grave, involve imagery from one input (the source input of death and grave digging), but the causal structure of the other input, in which the person is unwittingly contributing to his own future failure. Although the mapping might seem to draw an analogy between the grave digger and the fool, in fact digging a grave doesn’t cause anything (other than the grave itself), that might be mapped onto the grave-digger’s failure. Even in abhorrent instances such as that described in a 1995 Associated Press blurb (HEADLINE: YOUTH KILLED WITH SHOVEL, BURIED IN HOLE HE HAD DUG; Crime: A man and a teen-ager who allegedly taunted the victim before beating him to death are arrested)[2], the digging itself does not lead to death.

We suggest that the ready availability of literal meaning in idiom interpretation is no accident, as it stems from the import of conceptual structure in one or more of the input spaces in a conceptual integration network. Idiom interpretation requires the construction of a number of cognitive models, one of which corresponds to the source domain, and what would be dubbed a literal interpretation of the metaphoric expression. Moreover, the activation of conceptual structure from the source domain is not random, but seems to be limited to some metaphor-relevant aspects, with metaphor-irrelevant aspects being actively suppressed (Gernsbacher & Robertson, 1999). The context specificity of source domain activations may arise from inherent constraints on the alignment of structure between spaces in the network.

And so we find ourselves telling a story reminiscent of that told by linguists of old. While we reject a firm dichotomy between literal and nonliteral language, and argue that qualitatively similar processing operations underlie the comprehension of both sorts of meanings, our proposal is not too far removed from the old suggestion that readers construct a literal interpretation automatically as part of the parsing process. However, on the space structuring model, this grammatically cued meaning construction occurs more-or-less in parallel with the structuring of other spaces in the network. Consequently, parallel activation of meaning does not reflect a blind activation process to be followed by selection of the correct meaning. Rather, parallel activation is thought to reflect the construction of cognitive models in various spaces in the network. For this reason it is crucial for establishing the overall meaning, which involves comprehension of the relationship between the cognitive models in the source input, the target input, and the blended space. Continuity between literal and nonliteral language comprehension consists in the space structuring, mapping, and blending operations needed to construct literal and nonliteral meanings alike.

Acknowledgments

Many thanks to Raymond Gibbs for use of his psycholinguistics lab at University of California, Santa Cruz, and to research assistants Ashleigh Briggs, Jenny Lederer, Tracy Lee, and Annelise Casaubon-Smith.

References

Blasko, D.G., & Briihl, D. S. (1997). Reading and recall of metaphorical sentences: Effects of familiarity and context. Metaphor & Symbol, 12, 261-285.

Blasko, D.G., & Connine, C. M. (1993). Effects of familiarity and aptness on metaphor processing. Journal of Experimental Psychology: Learning, Memory, & Cognition,19, 295-308.

Cacciari, C., & Glucksberg, S. (1994). Understanding figurative language. In M.A. Gernsbacher (Ed.), Handbook of psycholinguistics (pp. 447-477). San Diego, CA: Academic Press.

Cacciari, C, & Glucksberg, S. (1995). Imaging idiomatic expressions: Literal or figurative meanings. In M. Everaert, & E. van der Linden (Eds.), Idioms: Structural and psychological perspectives (pp. 43-56). Hillsdale: Lawrence Erlbaum Associates.

Cacciari, C., & Tabossi, P. (1998). The comprehension of idioms. Journal of Memory & Language, 27, 668-683.

Coulson, S. (1996). The Menendez Brothers Virus: Analogical mapping in blended spaces. In A. Goldberg, (Ed.), Conceptual Structure, Discourse, and Language. Palo Alto, CA: CSLI.

Coulson, S. (1997). Semantic leaps: The role of frame-shifting and conceptual blending in meaning construction. San Diego, CA: University of California, San Diego, Ph. Dissertation.

Coulson, S. (2001, in press). Semantic Leaps: Frame-shifting and Conceptual Blending in Meaning Construction. New York and Cambridge: Cambridge University Press.

Coulson, S., King, J.W., & Kutas, M. (1998). ERPs and domain specificity: Beating a straw horse. Language and Cognitive Processes, 13, 653-672.

Coulson, S., & Van Petten, C. (2000). Conceptual integration and metaphor: An ERP study. Manuscript submitted for publication.

Fauconnier, G. (1994). Mental Spaces. New York and Cambridge: Cambridge University Press.

Fauconnier, G. (1997). Mappings in language and thought. New York, NY: Cambridge Press.

Fauconnier, G., & Sweetser, E. (1996). Spaces, Worlds, and Grammar. Chicago: University of Chicago Press.

Fauconnier, G., & Turner, M. (1998). Conceptual Integration Networks. Cognitive Science, 22, 133-187.

Frisson, S., & Pickering, M.J., (1999). The processing of metonymy: Evidence from eye movements. Journal of Experimental Psychology: Learning, Memory, & Cognition, 25, 1366-1383.

Frisson, S, & Pickering, M.J., (2000, in press). Metaphor & Symbol.

Gentner, D. (1983). Structure-mapping: A theoretical framework for analogy. Cognitive Science 7, 155-170.

Gentner, D, & Wolff, P. (1997). Alignment in the processing of metaphor. Journal of Memory & Language, 37, 331-355.

Gernsbacher, M. A. (1984). Resolving 20 years of inconsistent interactions between lexical familiarity and orthography, concreteness, and polysemy. Journal of Experimental Psychology: General, 113, 256-281.

Gernsbacher, M. A. and Robertson, R. (1999). The role of suppression in figurative language comprehension. Journal of Pragmatics 31 (12): 1619-1630.

Gibbs, R.W. (1993). Process and products in making sense of tropes. In A. Ortony (Ed.), Metaphor and Thought (pp.252-276). New York, NY: Cambridge University Press.

Gibbs, R. W. (1994). The poetics of mind : figurative thought, language, and understanding. New York, NY: Cambridge University Press.

Gibbs, R. W., & Gerrig, R. J. (1989). How context makes metaphor comprehension seem “special.”. Metaphor & Symbolic Activity, 4,145-158.

Gibbs, R.W., & Matlock, T. (in press). Looking for metaphor in all the right ways. Teoria et historia scientiarum: An interdisciplinary journal for interdisciplinary studies.

Gibbs, R.W., & Nascimento, S. (1993). How we talk when we talk about love: Metaphorical concepts and understanding love poetry. In R. Kreuz & M. MacNulty (Eds.), Empirical and aesthetic approaches to literature. Norwood, NJ: Ablex.

Giora, R. (1997). Understanding figurative and literal language: The graded salience hypothesis (Psycholinguistics). Cognitive Linguistics, 8, 183-206.

Giora, R., & Fein, O. (1999). On understanding familiar and less-familiar figurative language. Journal of Pragmatics, 31, 1601-1618.

Glucksberg, S. (1991). Beyond literal meanings: The psychology of allusion. Psychological Science, 2, 146-152.

Glucksberg, S. (1998). Understanding metaphors. Current Directions in Psychological Science, 7, 39-43.

Glucksberg, S., Gildea, P., & Bookin, H.B. (1982). On understanding nonliteral speech: Can people ignore metaphors? Journal of Verbal Learning & Verbal Behavior, 21, 85-98.

Glucksberg, S., & Keysar, B. (1993). How metaphors work. Metaphor and Thought (pp. 401-424). New York, NY: Cambridge University Press.

Grady, J., T., Oakley, & S. Coulson. (1999). Conceptual Blending and Metaphor. In R. Gibbs & G.J. Steen (Eds.), Metaphor in Cognitive Linguistics (pp. 101–124). Amsterdam & Philadelphia: John Benjamins.

Grice, H.P. (1975). Logic and conversation. In P. Cole and J.L. Morgan (Eds.), Syntax and Semantics, Vol. 3, Speech acts. New York: Academic Press.

Inhoff, A., Lima, S., & Carrol, P. (1984). Contextual effects on metaphor comprehension in reading. Memory and Cognition, 12, 558-567.

Keysar, B. (1989). On the functional equivalence of literal and metaphorical interpretations in discourse. Journal of Memory & Language, 28, 375-385.

Kutas, M., & Hillyard, S. (1980). Reading senseless sentences: Brain potentials reflect semantic incongruity. Science, 207, 203-205.

Kutas, M., Federmeier, K.D., Coulson, S., King, J.W., & Muente, T. (2000). Language. In J. T. Cacioppo, & L. G. Tassinary, (Eds.), Handbook of psychophysiology (2nd ed.). (pp571-601). New York, NY: Cambridge University Press.

Lakoff, G. (1993). The contemporary theory of metaphor. In A. Ortony (Ed.), Metaphor and thought (pp. 202-251), Second Edition. Cambridge: Cambridge University Press.

Lakoff, G., & Johnson, M. (1980). Metaphors we live by. Chicago, IL: Chicago University Press.

Lakoff, G., & Turner, M. (1989). More than cool reason. Chicago, IL: Chicago University Press.

Landauer, T.K., Foltz, P.W., & Laham, D. (1998). Introduction to Latent Semantic Analysis. Discourse Processes, 25, 259-284.

Landauer, T.K., & Dumais, S.T. (1997). A solution to Plato’s problem: The Latent Semantic Analysis theory of the acquisition, induction, and representation of knowledge. Psychological Review,104, 211-240.

McElree, B., & Nordlie, J. (1999). Literal and figurative interpretations are computed in equal time. Psychonomic Bulletin & Review, 6, 486-494.

Ortony, A., Schallert, D., Reynolds, R., & Antos, S. (1978). Interpreting metaphors and idioms: Some effects of context on comprehension. Journal of Verbal Learning and Verbal Behavior 24, 569-594.

Pynte, J., Besson, M., Robichon, F., & Poli, J. (1996). The time-course of metaphor comprehension: An event-related potential study. Brain and Language, 55, 293-316.

Rugg, M.D., & Coles, M.G.H. (1995). The ERP and cognitive psychology: Conceptual issues. In M.D. Rugg, & M.G.H. Coles (Eds), Electrophysiology of mind: Event-related brain potentials and cognition (pp. p. 27-39). Oxford, England: Oxford University Press.

Searle, J.R. (1979). Expression and Meaning: Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press.

Sweetser, E. (1990). From etymology to pragmatics: Metaphorical and cultural aspects of semantic structure. Cambridge, Massachusetts: Cambridge University Press.

Turner, M. (1991). Reading minds: The study of English in the age of cognitive science. Princeton, N.J.: Princeton University Press.

Turner, M. (1996). The literary mind. New York: Oxford University Press.

Turner & Fauconnier. (2000). Metaphor, metonymy, and binding. In A. Barcelona (Ed.), Metonymy and Metaphor at the Crossroads (pp. ). New York: Walter de

Gruyter.

Wolff, P, & Gentner, D. (2000). Evidence for role-neutral initial processing of metaphors. Journal of Experimental Psychology: Learning, Memory, & Cognition, 26, 529-541.

Table 1

Spaces in Conceptual Integration Network for the Unsinkable Titanic

__________________________________________________

Generic Space Input Input Blended

__________________________________________________

object ship movie ship

course voyage run voyage

outcome sunk wins-oscars sink

__________________________________________

Table 2

Examples of Some Features Generated with Metaphoric, Literal-Mapping, and Literal Contexts

offshoot

met You might think ambition is a productive emotion, but jealousy is often its offshoot

Unique features: DOWN-SIDE, UNWANTED, MOTIVATION, REASON

lit-map The way those two trees have grown together, the left one looks like an offshoot

Unique features: FORK, CONNECTION, LEAN

lit

The Rockies are the major mountain range around here, this one is just an offshoot

Unique features: SMALL, EXTRA, ADDITION, SUBSIDIARY, RANDOM

Shared features (appear in 3 different contexts): BRANCH, GROW

meteor

met Spectacular and short-lived, the right mix of gin and vermouth is a meteor

Unique features INTOXICATING, STRONG

lit-map Not well versed in astronomy, she mistakenly thought the comet was a meteor

Unique features FALLING FROM THE SKY, FLASH, DANGER, BALL

lit She looked up into the night sky and happened to see a meteor

Unique features DISTANT, UNIVERSE, EXPANSIVE

Shared features: FAST, SHOWER, ROCK, BRIGHT, SHOOTING

reststop

met She said it was serious but her relationship with him was just a reststop

Unique features NOTHING SERIOUS, WAITING, IN BETWEEN, REBOUND

lit-map Looking at the photo closely he realized the campground was actually a reststop

Unique features PARK, PLACE ALONG THE ROAD, OPEN, RECREATION

lit After tracking him for days, the police finally cornered the fugitive

Unique features INTERSTATE, STOPOVER, PITSTOP

Shared features: BREAK, BATHROOM, RESTROOM, RELAX, HIGHWAY

cattle

met Blindly following orders, those cult members were cattle

Unique features BLIND, STUPIDITY, DEPENDENT, UNTHINKING, DEATH

lit-map He mistook the herd of gazelles for cattle

Unique features horns, wild, goats, DOMESTICATED

lit We grew some corn for ourselves but more of it for the cattle

Unique features FOOD, CHEWING, and VARIOUS STOMACHS

Shared features: COW, ANIMALS, MEAT

FIGURE CAPTIONS

Figure 1. Classic N400 Effect. The solid line shows the event-related brain response (ERP) from one electrode site for processing words that were highly expected in the context. The dashed line shows the ERP elicited by words that were unexpected in the context.

Figure 2. Percentages for unique features generated per context type. Error bars represent the standard error of the mean.

[1] Of course, the underspecification model could be resuscitated if it were found that brain activity underlying the N400 is correlated to measures of total reading time rather than to the first fixation measure used by Frisson & Pickering (see this volume for review). At present first fixation is the best sign of immediate processing difficulty in eye tracking studies of visual language comprehension. N400 is the best sign of immediate difficulty of lexical integration in the ERP. However, the exact relationship between the two measures is currently unkown.

[2] Thanks to Todd Oakley for bringing this (albeit gruesome) example to our attention.

2002. Cognitive Linguistics.

Surface Generalizations: an alternative to alternations[1]

Adele E. Goldberg

University of Illinois

Abstract

Since the earliest days of generative grammar, there has existed a strong tendency to consider one argument structure construction in relation to a particular rough paraphrase. Initially this was a result of the emphasis on transformations that derived one pattern from another. While today there exist many non-derivational theories for which this motivation no longer exists, the traditional outlook has not completely lost its grip, as can be seen from continuing focus on partial or incomplete generalizations such as the “dative” construction or the “locative” alternation. This paper argues that it is profitable to look beyond alternations and to consider each surface pattern on its own terms. Differences among instances of the same surface pattern are often most naturally attributed directly to the different verbs and arguments involved.

1. The Surface Generalization Hypothesis

Many theoretical approaches today eschew the need for any kind of transformation or derivation (e.g., Bresnan 1982; Bresnan 1994; Fillmore, Kay, and Michaelis in progress; Lakoff 1987; Langacker 1987a; Langacker 1991; Pollard and Sag 1987). A compelling reason to avoid positing derivations in favor of an emphasis on surface form is simply that there are typically powerful generalizations surrounding particular surface forms that are more broad than those captured by derivations or transformations. We refer to these broader generalizations as Surface Generalizations. The present paper focuses on the domain of argument structure; the surface formal and semantic/pragmatic generalizations in this domain are captured by Argument Structure Constructions: pairings of form and function that are used to express basic clauses.

In this paper, several case studies are considered including the “dative” construction and the “locative alternation.” It is argued that traditional divisions under-represent the generalizations that exist. We address the question of how to account for paraphrase relations, as well as how to account for various differences between instances of the same argument structure construction in section 5. In this section we review an important historical precedent for the form of argument made here.

Despite being the most influential architect of transformations and later, derivations, Chomsky (1970) put forward one of the most well-known and widely accepted arguments against deriving one subset of data from another. His argument was based on Surface Generalizations. In particular, he demonstrated that NPs based on “derived” nouns (i.e., nouns that have verbal counterparts) have exactly the syntax of NPs based on underived nouns. In particular they both have the same internal and external syntax. Both types occur with the full array of determiners, often pluralize, and take complements marked with of. Both types can appear as the subject of passives or can be distantly instantiated by a question word. To avoid an account in which this is mere coincidence, Chomsky reasoned, we need to recognize that both types are base-generated as nouns instead of attempting to derive certain NPs from clausal counterparts (Lees 1960). With Williams (1991), we might call this the “target syntax argument”: it is preferable to generate A directly instead of deriving it from C if there exists a pattern B that has the same target syntax as A and is clearly not derived from C.

Williams (1991) makes a parallel “target semantics argument.” He observes that the meanings of NPs based on underived nouns fall into the same set of categories as the meanings of NPs based on “derived” nouns. For example, extent, temporal duration, and evaluative states can be predicated of both “derived” and underived nouns (1991:584):

Extent:

(1) a. The destruction of the city was complete. Potentially derived

b. The carnage was complete. Underived

Temporal duration

(2) a. The destruction of the city took four hours. Potentially derived

b. The war took for hours. Underived

Evaluative state

(3) a. The destruction of the city was horrible. Potentially derived

b. The war was horrible. Underived

(not just the fact of the war, but the way the war was)

At the same time, Williams argues, the range of NP meanings is distinct from the range of S meanings, as seen in examples (4)-(6) (1991: 585)[2]:

Extent

(4) *That the city was destroyed was complete.

Temporal duration

(5) *That the city was destroyed took four hours.

Evaluative state

(6) *That the city was destroyed was horrible. (just the fact that the city was destroyed, not the way it was destroyed)

In short, given that the syntax and semantics of derived nouns are like those of underived nouns, and unlike the syntax and semantics of clauses, it is clearly simpler to allow the nouns to be base-generated as nouns as opposed to deriving them from clause structures.

Beyond target syntax and target semantics arguments are what are referred to below as “input” syntax and semantics arguments. In particular, it is preferable to avoid deriving A from C if there exists a pattern B that has the same syntax and semantics as C and yet cannot serve as input from which to derive A.

The arguments put forth by Chomsky (1970) (and Williams 1991) have been robust. For more than three decades, the field has resisted the temptation to derive deverbal NPs from clauses. What is less widely recognized is that parallel arguments hold in the domain of argument structure. These arguments support the idea that each argument structure pattern is best analyzed on its own terms, without relying on explicit or implicit reference to a possible alternative paraphrase. It is argued that such reliance effectively puts blinders on, and limits a theory’s ability to state the full extent of the relevant generalizations.

We might label the hypothesis that the target syntax and target semantics arguments and the input syntax and semantics arguments hold in general for argument structure patterns, the Surface Generalization Hypothesis.

Surface Generalization Hypothesis: there are typically broader syntactic and semantic generalizations associated with a surface argument structure form than exist between the same surface form and a distinct form that it is hypothesized to be syntactically or semantically derived from.

Support for the Surface Generalization Hypothesis provides substantial motivation for the assumption that the syntax of argument structure should be represented without recourse to derivations. Perhaps more relevantly for the present audience, it also suggests that it is possible to overplay the importance of alternative forms (paraphrases).

In section 2 the ditransitive construction is discussed. Section 3 analyzes the dative paraphrase and the benefactive paraphrase as instances of broader argument structure constructions. Section 4 focuses on generalizations beyond the “load/spray” alternation. In section 5 we focus on the role of individual verbs and argue that they serve to capture what is shared between members of an alternation; it is also argued that attention to individual verbs allows us to motivate distinctions among instances of what are argued to involve the same general argument structure construction. Section 6 clarifies what is intended by “surface form;” in this section it is suggested that argument structure constructions in English do not specify word order but instead are better captured by a set of grammatical relations together with the corresponding semantic interpretation.

2. The Ditransitive Construction

Many generative theories derive the two ditransitive or double object expressions in (7) from distinct input expressions on the left, which correspond to their rough paraphrases (Baker 1988; Larson 1988):

(7) a. Mina bought a book for Mel.à Mina bought Mel a book.

b. Mina sent a book to Mel. à Mina sent Mel a book.

Even certain constructional approaches treat the two examples on the right at instance of two independent constructions (e.g., Jackendoff 1990; Kay Ms-2001). However, both instances of the ditransitive share many properties with each other and differ systematically from their paraphrases (see also Langacker 1991; Oehrle 1975). That is, there are good reasons to group the two “outputs” together as distinct from the “inputs” as follows:

Mina bought a book for Mel. Mina bought him a book.

Mina sent a book to Mel. Mina sent Mel a book.

Obvious similarities between the two ditransitive expressions begin with their shared surface form; in its simple active form, the ditransitive involves an active verb followed by two NPs. Both ditransitives readily allow the theme argument to be distantly instantiated, for example as a question word:

(8) a. What did Mina buy Mel?

b. What did Mina take Mel?

In both cases questioning the recipient argument is less acceptable:

(9) a. ??Who did Mina buy a book?

b. ??Who did Mina take a book?

Both paraphrases, on the other hand, allow either the recipient or theme argument to be questioned with equal ease:

(10) a. Who did Mina buy a book for?

b. Who did Mina take a book to?

(11) a. What did Mina buy for Mel?

b. What did Mina take to Mel?

The abilility to form passive has been claimed to differentiate ditransitives into two types; it has been claimed that that only those with paraphrases involving to can be passivized (Fillmore 1965; Kay Ms-2001). While it may be true that ditransitives that have paraphrases with to show a statistical tendency to passivize more easily than those that have paraphrases with for, the generalization is far from clear cut as many have observed (see Culicover and Wexler 1973; Erteschik-Shir 1979; Oehrle 1975). For example, the following examples appear to be equally acceptable (or if anything, a. is more acceptable than b. despite the fact that only b. is paraphrasable with t)::

(12) a. Mel was cooked a fine dinner by the new chef. (cf. The new chef cooked a fine dinner for Mel)

b. Mel was tossed a blanket by the babysitter. (cf. The babysitter tossed a blanket to Mel.)

There are additional ways in which all ditransitives pattern alike. Adverbs may not separate the two NP arguments in ditransitives (13), while they can separate the direct object from the for prepositional phrase as in (14a) and to some extent can separate the direct object from the “to” prepositional phrase as in (14b):

(13) a. *Mina bought Mel yesterday a book.

b. *Mina sent Mel yesterday a book.

(14) a. Mina bought a book yesterday for Mel.

b. ?Mina sent a book yesterday to Mel.

Neither type of ditransitive expression allows the theme argument to be the third person singular it (Green 1974; Oehrle 1976):

(15) a. ??Mina sent Mel it.

b. ??Mina bought Mel it.

This restriction does not hold of either prepositional paraphrase:

(16) Mina sent it to Mel.

(17) Mina bought it for Mel.

Beyond, and often behind the similarities of the surface form of a construction there lie shared functional similarities. In the case of the ditransitive, all instances share identical information theoretical constraints and have closely related semantics. [3] That is, information structure properties group ditransitives together as a class. In both so-called to and for ditransitives, for example, the recipient argument tends to be shorter in length and already given in the discourse, as compared to either prepositional paraphrase (Arnold et al. 2000; Erteschik-Shir 1979; Thompson 1990).

Semantically, both so-called for ditransitives and so-called to ditransitives require that the recipient argument be construed to be animate (Green 1974; Oehrle 1975; Partee 1965/1979):[4]

(18) a. ??Mina sent that place a box.

b. ??Mina bought that place a box.

This restriction is again not relevant to either prepositional paraphrase:

(19) a. Mina sent a box to that place.

b. Mina bought a box for that place.

More generally, the particular meaning associated with the ditransitive evokes the notion of “giving” in various ways, depending on the verb class involved. This is in contrast to paraphrases with for. For example, while (20) can be used to mean that Mina bought a book for a third party because Mel was too busy to buy it himself, (21a) can only mean that Mina intended to give Mel the book (Green 1974; Oehrle 1976; Goldberg 1992). The semantics of giving is likewise apparent in (21b):

For paraphrase:

(20) Mina bought a book for Mel.

(the book could be intended for Mel’s mother, bought by Mina because Mel was too busy to buy it)

Ditransitives:

(21) a. Mina bought Mel a book. (Mina intends to give Mel the book)

b. Mina sent Mel a book. (Mina again intends to give Mel the book)

Other interpretations for the ditransitive can also be systematically related to the notion of giving, in that they may imply that the transfer will occur if certain satisfaction conditions evoked by the main verb occur (22a), that transfer will not occur (22b), or that the antonymic relation of giving, that of taking away occurs (22c).[5]

(22) a. Mina guaranteed/offered Mel a book. (If the guarantee or offer is satisfied, Mel will receive a book)

b. Mina refused Mel a book. (Mina caused Mel not to receive a book)

c. Mina cost Mel his job. (Mina causes Mel to lose his job).

It has been suggested that the existence of variable meanings undercuts the claim of a unified construction (Nakajima 2002). The criticism stems from the belief that the concepts of, for example, giving, not giving, and taking away cannot naturally be classed together. However, it is clear that both the negation and the antonym of a particular concept are closely associated with that concept. For example, a concept and its antonym typically serve as strong associates for one another in psycholinguistic studies (Meyer and Schvaneveldt 1971): e.g., hot primes cold, high primes low, and giving primes taking away. Negated sentences typically presuppose that the corresponding positive assertion has been asserted or might be believed in the particular context of use (Givón 1979). In this way we can see that giving, not giving, and taking away are in fact closely associated concepts.

Thus we see that ditransitives expressions pattern alike on a number of syntactic and semantic dimensions regardless of their potential paraphrases. It seems that the only thing that the respective paraphrases share with the ditransitives is the quite rough paraphrase relations themselves. There is little empirical motivation to decree that ditransitives must be derived from prepositional paraphrases nor that ditransitives that admit of distinct paraphrases must be treated as themselves more than minimal variants of each other. The robust generalizations are surface generalizations.

3. The Caused Motion and Benefactive Constructions

Beyond target syntax and target semantics arguments are input syntax and semantics arguments: it is preferable to avoid deriving A from C if there exists a pattern B that has the same target syntax and semantics as C and yet cannot serve as input from which to derive A. By widening our focus beyond those expressions that may serve as paraphrases of ditransitives, we see that each paraphrase expression itself is a small part of a much broader generalization. For example, although only (23a) can be paraphrased by a ditransitive, it patterns together with (23b, c, d) both syntactically and semantically; in fact, all of the expressions in (23) can be captured by a single “caused-motion” construction (Goldberg 1995; cf. also Pinker 1989).

(23) a. Mina sent a book to Mel.

b. Mina sent a book to Chicago.

c. Mina sent a book toward the front of the room.

d. Mina sent a book through the metal detector.

Similar extensions of meaning that we saw above for the ditransitive likewise exist in the case of the caused motion construction, even though the verb classes involved are distinct:

(24) a. Mina coaxed Mel into the room. (if coaxing is successful, Mel moves into the room)

b. Mina helped Mel into the room. (Mina helps Mel move into the room)

c. Mina blocked Mel out of the room. (Mina causes Mel not to move into the room)

The for paraphrase of certain ditransitives (e.g., 25a) patterns together with (25b, c) syntactically and semantically; each are instances of a transitive construction together with a benefactive adjunct construction.

(25)a. Mina sent a book for Mel.

b. Mina sent a book for the library.

c. Mina sent a book for her mother’s sake.

An objection might be raised against the proposal that all for-benefactive phrases should be treated as a natural class. It might be argued that because more than one can co-occur, they cannot play the same role in the sentence:

(26) Mina sent a book for Mel for her mother’s sake.

That is, Fillmore (1968) long ago observed that only one semantic role of each type may occur in a single clause. We do not find two distinct agents or patients co-occurring in a single clause:

(27) *Bob melted the butter by Paul.

(28) *The butter was melted the ice.

But Fillmore’s constraint only holds of certain semantic roles, namely those that can be identified as arguments. Adjuncts can freely be added as long as they do not imply a semantic contradiction; in particular they must be construed to have concentric semantic scope such that one more narrowly specifies another. Consider the following sentence with multiple temporal adjuncts:

(29) Mina met Bob in the morning yesterday at 11 o’clock.

Notice the hour (here 11 o’clock) must occur within the part of the day (morning) which is in turn within the day (yesterday). It cannot be claimed that the temporal phrases must be interpreted syntactically as a single complex adjunct because they need not be continuous:

(30)a. Yesterday Mina met Bob in the morning at 11 o’clock.

b. At 11 o’clock in the morning Mina met Bob yesterday.

c. Yesterday Mina met Bob in the morning by the beach at 11 o’clock

More than one locational adjunct can likewise appear in a single clause:

(31) Mina met Mel on the beach in California near the boardwalk.

Again, the locational adjuncts need not be continuous and therefore are distinct adjunct phrases:

(32)a. In California, Mina met Mel on the beach near the boardwalk.

b. Near the boardwalk Mina met Mel on the beach in California.

c. On the beach in California Mina met Mel near the boardwalk.

Thus the fact that more than one for phrase can appear does not necessarily undermine the argument that each of those in (25) is a benefactive phrase. The for phrases are all headed by the preposition for and they all encode a benefactive relation. Moreover each functions as an adjunct. The shared syntax and semantics of these phrases argue for treating them alike.

It should be made clear that we are not claiming that all for phrases encode benefactives. Clearly there are other uses of the preposition for in English which may not be related, for example, those in (33). Prepositions are typically highly polysemous and sometimes ambiguous (see Brugman 1988; Lakoff 1987; Lindner 1981).

(33)a. The statue stood for three hours.

b. He exchanged the socks for a belt.

That is, there do exist instances of constructional ambiguity: a single surface form having unrelated meanings.[6] It must be emphasized that it is not being claimed that meaning is simply read off surface form. What is being suggested here is simply that by putting aside rough paraphrases and considering all instances with a formal and semantic similarity, broader generalizations can be attained. In order to identify which argument structure construction is involved in cases of constructional ambiguity, attention must be paid to individual verb classes. In fact, in order to arrive at a full interpretation of any clause, the meaning of the main verb and the individual arguments must be taken into account. This is discussed in more detail in section 5. In cases such as those in (25), what is being proposed is simply that if a constituent looks like a benefactive phrase and acts like a benefactive phrase, then there is no reason to be shy about calling it a benefactive phrase.

Thus the input syntax and semantics arguments strengthen the case against deriving ditransitives from their corresponding paraphrases which have prepositional arguments. The formal patterns involved should be viewed as constructions on their own terms—the ditransitive, the caused motion, the simple transitive and the benefactive adjunct constructions:

Mina bought Mel a book. Ditransitive Construction

Mina sent Mel a book.

Mina sent a book to Mel. Caused-Motion Construction

Mina bought a book for Mel. Transitive construction + benefactive adjunct construction

Each of these constructions can be seen to be much more general than is often recognized when only instances that alternate in certain ways are considered. A more representative array of instances of each construction is provided below:

Ditransitive

(37a) Mina bought Mel a book.

Mina sent Mel a book.

Mina gave Mel a headache.

Mina fixed me a sandwich.

Mina guaranteed/offered Mel a book.

Mina refused Mel a book.

Mina cost Mel his job.

Caused-Motion Construction

(38a) Mina sent a book to Mel.

b. Mina sent a book to Chicago.

c. Mina tossed a book toward the front of the room.

d. Mina put a book through the metal detector.

e. Mina coaxed Mel into the room.

f. Mina helped Mel into the room.

g. Mina blocked Mel out of the room.

Transitive construction + benefactive adjunct construction

(39a) Mina sent a book for Mel.

b. Mina sent a book for the library.

c. Mina sent a book for her mother’s sake.

4. Load/Spray

Similar arguments can be made for other types of argument structure patterns that are often only considered in terms of alternations (Anderson 1971; Fraser 1971; Hook 1983; Rappaport and Levin 1988). Consider the following examples in (40) and (41).

(40) Pat loaded the wagon with the hay.

(41) Pat loaded the hay onto the wagon.

It has been suggested that the with variant is derived from the into variant (e.g., Rappaport and Levin 1988). Let us consider the “input” syntax and semantics first. The “into” variant can be seen to be an instance of the much broader caused-motion construction already discussed. That is, each of the examples in (42) shares the same surface syntax: each has a DO and prepositional oblique phrase. The semantics are closely related as well; in each case the subject argument serves to cause the motion of the DO argument along the path or to the location specified by the oblique argument:

(42) a. Pat loaded the hay onto the wagon.

b. Pat put the hay on the wagon.

c. Pat shoveled the hay into the wagon.

The b and c forms of (42) cannot serve as input to any locative alternation as can be seen in the ill-formedness of the following examples:[7]

(43) b. *Pat put the wagon with hay.

c. *Pat shoveled the wagon with the hay.

We thus see that the input syntax and semantics arguments hold for the into variant of the so-called locative alternation. We now turn to the putative “output” syntax. Making the argument that the with variant is an instance of a broader generalization is somewhat more controversial than the other cases discussed so far primarily because with has a remarkably wide range of uses, a point we return to below. Consider just a sampling of various uses of with:

(44) a. Elena traveled with Maya.

b. Elena traveled with a hat on.

c. Aliza traveled with great enthusiasm.

d. People associate one variant with another.

e. Be sure to mix the butter with sugar.

f. The foundation provided the school with funding.

g. Pat loaded the wagon with hay.

h. The garden swarmed with bees.

i. The detective entered the room with a key.

j. Pat broke the window with a hammer.

k. Pat watched the bear with a telescope.

One would have to be quite an ardent lumper to try to class all of these uses of with under a single sense. Again, this is not the claim of the present paper: we do not deny the existence of constructional ambiguity. It is suggested, however, that it is important not to assume massive ambiguity without seeking out broader surface generalizations.

Consider just the following examples that have been independently classified as instances of the “locative” construction by Pinker (1989) and Levin (1993):

(45) a. Pat loaded the wagon with the hay.

b. Pat sprayed the wall with paint.

c. They covered the wall with posters.

d. Pat adorned the tree with lights.

e. They tiled their bathroom with blue tiles from Mexico.

f. They stained the wood with an all-weather protector.

g. He speckled the canvas with dots.

h. He wrapped the present with tin foil.

It is possible to make a case that the examples in (45) are all licensed by the combination of two constructions: a causative construction and an independent construction headed by with. By recognizing that the transitive syntax and semantics in each of the examples in (45a-h) is licensed by a causative construction, we account for the well-known fact that the DO in these examples is necessarily interpreted as affected in some way; e.g, the truck must be interpreted to be full or otherwise affected in Pat loaded the truck (with hay). The same is not true for Pat loaded hay onto the wagon (Anderson 1971), which only entails that some hay is put on the wagon. By acknowledging that the examples in (45) are causative, the affected status of the DO is rendered completely non-mysterious and requires no ad hoc stipulation. This proposal has also been put forward by Rappaport and Levin (1988) and Gropen et al. (1991).

Iwata (2002), however, argues against a causative analysis of examples like She loaded the wagon with hay. He argues that a causal analysis should predict that the examples should necessarily be telic, which they are not:

(46) He sprayed the lawn with water for hours/in an hour.

However, others have noted that aspectual status is largely independent of causal status (Jackendoff 1996; Van Valin and LaPolla 1997). For example, clear instances of causal predications can also be used with either an atelic or telic interpretation:

(47) a. He mowed the lawn for hours/in an hour.

b. He cut the fabric for hours/in an hour.

c. He broke the walnuts for hours/in an hour.

A second argument Iwata suggests is that the load class of verbs differs from lexical causatives such as cut, destroy, kill, break, open, melt in that the former specifies the manner in which the resulting change is achieved. However, it is not obvious that load specifies the manner of the process any more than melt does. Things can be loaded manually or by machine, quickly or slowly; load only requires that the entities be put somewhere with substantial but limited space. Likewise while things can be melted by the sun or the stove, melt does specify that the change of state is caused by an application of heat. Other causative verbs are even more specific about the manner of the process including strangle (to kill by using hands around the neck), murder (to kill intentionally), and bludgeon (to injure by using forceful blows of a blunt object).

Moreover, while we claim that the with variant is causal, it does not follow that the verbs that appear in it are necessarily interpreted causally when they appear in other constructions. The verbs need only be compatible with a causal interpretation. When the verbs are not used causally (as in the into/onto variant), they do not imply that the location argument undergoes a change of state. In fact, because load/spray verbs are not always used as causative verbs, they by necessity must specify something besides a resultant endstate. Thus the observation that alternating verbs of the load type necessarily encode more than causation is expected. We return to clarify the distinction and relation between verbs and constructions in section 5.

Adopting then the idea that the examples in (47a-h) admit of a causal analysis, notice none of the examples in (47c-h) permit the alternation typically discussed as being relevant to load and spray (see also Pinker 1989):

(46) a. Pat loaded the hay onto the wagon.

b. Pat sprayed paint onto the wall.

c. *They covered posters onto the wall.

d. *Pat adorned lights onto the tree.

e. *They tiled blue tiles from Mexico onto their bathroom

f. *They stained an all-weather protector onto the wood.

g. *He speckled dots onto the canvas.

h. *He wrapped tin foil onto the present.

Thus in accord with the target syntax argument, it is preferable to generate (45a,b) directly instead of deriving them from (46a,b) due to the fact that there exist (45c-h) that have parallel syntax and semantics and cannot be derived from 46(c-h).

Turning our attention to the with phrase, it can be observed that while there is likely no monosymous sense for the preposition, there are a number of reasons to conclude that the with in the expressions in (47) is related to the instrumental adjunct with in (48). Formally it shares the same preposition with prototypical instrumentals, not only in English, but also in a number of other languages (Rappaport and Levin 1988). Semantically, the entity encoded by the with phrase is in both cases manipulated by the subject argument and serves to effect the change of state entailed by the sentence. In both cases, the argument of with serves as an intermediary in the causal chain.

(47) Pat loaded the wagon with hay.

(48) Pat broke the window with a hammer.

Below we repeat the examples in (45a-h) as (49a-h) and add to them examples (49i-m). It is difficult to draw a clear division in this set, exhaustively dividing them into clear instrumentals and clear non-instrumentals. In some cases (e.g., 49m) it is an independent tool that makes contact with the patient argument; in other cases (e.g., 49a) the entity serves to encode an argument of the verb that specifies something that is moved onto the patient. However, in still other cases, the argument bears both relations simultaneously (e.g., 49h-k). Again, in all cases, the with phrase encodes an entity that serves as an intermediary between agent and patient in a causal chain.

(49) a. Pat loaded the wagon with the hay.

b. Pat sprayed the wall with paint.

c. They covered the wall with posters.

d. Pat adorned the tree with lights.

e. They tiled their bathroom with blue tiles from Mexico.

f. They stained the wood with an all-weather protector.

g. He speckled the canvas with dots.

h. He wrapped the present with tin foil.

i. She broke the fever with cool washcloths.

j. She warmed the child with a blanket.

k. She loosened the cap with hot water.

l. She loosened the cap with a spoon.

m. She broke the window with a hammer.

It might be argued that the with phrase in (47) is crucially distinct from the instrumental with because it can appear with an additional instrumental phrase as in (50) (see Van Valin and LaPolla 1997):

(50)Pat loaded the wagon with the hay with a pitchfork.

However, as was noted in the case of temporal and locative adjuncts, it is possible to add an additional with instrumental phrase to prototypical instrumental adjuncts. In this case, the syntactically more peripheral phrase is understood to have broader scope than the more internal phrase.[8]

(51)a. With a slingshot he broke the window with a rock.

b. The robot opened the door with a key with its robotic arm.

We are assuming that the additional with phrase heads an adjunct in these cases; we refine our understanding of arguments and adjuncts in section 5 below. Also in section 5, we acknowledge certain ways in which the with phrase of load patterns differently than the with phrase in certain other examples in (49a-m). It is argued that these distinctions naturally follow from lexical semantic differences in the verbs involved and do not necessitate treating the with phrases as instances of unrelated constructions.

Still, it could be that instead of appealing to the notion of an intermediary to capture what is shared by the with phrases in (49a-m), a better analysis of these uses of with would be one that appeals to the idea of grammatical chains (Heine 1992). One use of with may be closely related to another, that one to a third and so on; that does not necessarily imply that the first and third are themselves of the same type. To assume that they are would be to fall prey to the Slippery Slope Fallacy; e.g., while a child is much the same from one day to the next, it does not follow that there is no distinction between a boy and a man. While a full analysis of with is outside the scope of the present paper, it is argued here that the relationship among various uses of with deserves exploration and cannot be dismissed out of hand.

5. The role(s) of the verb

In this section, we address the question of how to account for the overlap in meaning in paraphrases and we address the question of why the overt interpretation of instances of the same construction may differ, and may allow distinct ranges of paraphrases. One key to these questions lies in the recognition that there is more to the interpretation of a clause than the argument structure construction used to express it. The overall interpretation is arrived at by integrating the argument structure construction with the main verb and various arguments, in light of the pragmatic context in which the clause is uttered.

There is a growing recognition that it is important to recognize a distinction between the frame semantics associated with a verb and the set of phrasal patterns or argument structure constructions that are available for expressing clauses (Gleitman et al. 1995; Goldberg 1992, 1995, to appear; Hovav and Levin 1998; Iwata 2000; Jackendoff 1997, 2002; Kay Ms-2001; Pinker 1994).

Following Goldberg (1992, 1995) the slots in the argument structure constructions are referred to as “argument roles.” That is, phrasal constructions that capture argument structure generalizations have argument roles associated with them; these often correspond roughly to traditional thematic roles such as agent, patient, instrument, source, theme, location, etc. At the same time, because they are defined in terms of the semantic requirements of particular constructions, argument roles in this framework are more specific and numerous than traditional thematic roles (see also Jackendoff 1990, 2002).

Argument roles capture surface generalizations over individual verbs’ participant roles.

That is, each distinct sense of a verb is conventionally associated with rich frame semantic meaning that in part specifies certain participant roles: the number and type of slots that are associated with a given sense of a verb. A subset of those roles, namely those roles which are lexically profiled, are obligatorily expressed, or, if unexpressed, must receive a definite interpretation.[9] Lexical profiling, following the general spirit of Langacker (1987, 1991), is designed to indicate which participant roles associated with a verb’s meaning are obligatorily accessed, functioning as focal points within the scene, achieving a special degree of prominence. Fillmore (1977) similarly notes that certain participant roles are obligatorily “brought into perspective” achieving a certain degree of “salience.” The notion of lexical profiling is intended to be a semantic one: it is a stable aspect of a word’s meaning, and can differentiate the meaning difference between lexical items—cf. buy vs sell (Fillmore 1977) or rob vs steal (Goldberg 1995). Participant roles may be highly specific and are often unique to a particular verb’s meaning; they therefore naturally capture traditional selectional restrictions.

Two general principles can be understood to constrain the ways in which the participant roles of a verb and the argument roles of a construction can be put into correspondence or “fused”: the Semantic Coherence Principle and the Correspondence Principle (Goldberg 1995, to appear).

The Semantic Coherence Principle ensures that the participant role of the verb and the argument role of the construction must be semantically compatible. In particular, the more specific participant role of the verb must be construable as an instance of the more general argument role. General categorization processes are responsible for this categorization task and it is always operative. This principle follows from the idea that argument structure constructions are learned by generalizing over the semantics of instances of the pattern used with particular verbs (e.g., Tomasello 1992, 2000; Goldberg 1999).

As is the case with lexical items, only certain argument roles are profiled. In the case of simple sentences, only roles that are realized as Subj, Obj, or the second object in ditransitives are considered profiled. These are the same grammatical relations that receive a special status in most theories as the set of “terms” which correspond to “core,” “nuclear” or “direct” arguments. Roles encoded by the subject, object or second object grammatical relations are afforded a high degree of discourse prominence, being either topical or focal in the discourse (see Keenan 1976, 1984; Comrie 1984; Fillmore 1977, Langacker 1987 for arguments to this effect.). Specifically the Correspondence Principle states that profiled participant roles of the verb must be encoded by profiled argument roles of the construction, with the exception that if a verb has three profiled roles, one can be represented by an unprofiled argument role (and realized as an oblique argument). The Correspondence Principle is a default principle.

The intuition behind the Correspondence Principle is that lexical semantics and discourse pragmatics are in general aligned. That is, the participants that are highly relevant to a verb’s meaning (the profiled participant roles) are likely to be the ones that are relevant or important to the discourse, since this particular verb was chosen from among other lexical alternatives. In particular, the Correspondence Principle requires that the semantically salient profiled participant roles are encoded by grammatical relations that provide them a sufficient degree of discourse prominence: i.e. by profiled argument roles. As a default principle, the Correspondence Principle is overridden by particular constructions that specify that a particular argument be deemphasized and expressed by an oblique or not at all. Passive, for example is a construction that overrides the Correspondence Principle and insures that a normally profiled role (e.g., agent) be optionally expressed in an oblique by phrase. See Goldberg (to appear) for discussion of other constructions that serve to override the Correspondence Principle.

5.1. Accounting for paraphrase relations

We are now in a position to address the question of how the overlap in meaning between alternants is accounted for. The shared meaning can be attributed directly to the shared verb involved. That is, the verb evokes the same frame semantic scene and the same profiled participant roles. For example if we assign the participant roles of load the labels loader, loaded-theme and container, we can see that these roles line up with the roles in the caused motion construction and causative + with constructions as follows:

(52) Caused-motion (e.g., Pat loaded the hay onto the truck)

CAUSE-MOVE (cause theme path/location)

Load (loader loaded-theme container)

(53) Causative construction + with construction (e.g., Pat loaded the truck with hay)

CAUSE (cause patient) + INTERMEDIARY(instrument)

Load (loader container loaded-theme)

All three of load’s roles are profiled. This includes the loaded-theme role even though that role is optional. When optional, it receives a definite interpretation as indicated by the strangeness of the following mini-conversation (see Fillmore 1986 for tests to distinguish definite from indefinite omission):

(54) She loaded the trucks. #I wonder what she loaded onto the trucks.

Because all three roles are profiled, one of the roles may be expressed as an oblique argument, in accordance with the Correspondence Principle. The Semantic Coherence Principle insures that only semantically compatible roles may be fused. As indicated above, the loaded-theme role of load may either be construed to be a type of theme as in (52) or an intermediary as in (53). The container role can either be construed to be a path/location as in (52) or a patient role as in (53). Construing the verb’s roles as instances of different argument roles is what results in the different semantic construals of the two constructions.

On this view, there is no need to say that the with phrase itself designates a theme relation (cf. e.g., Jackendoff 1990). Instead, the fact that the hay is interpreted to be loaded onto the truck even in the with variant is attributed, not to the argument structure construction, but to the specifications of the verb load.

5.2. Arguments and Adjuncts

Recognizing that the verb has its own profiled participant roles that may be distinct from the argument roles associated with an argument structure construction allows us to recognize the following four possibilities:

Role of

argument structure constructionNot a role of argument structure construction

profiled/

obligatory

participant role

of verba) ARGUMENT

of verb and construction

He devoured the artichokes.

She gave him a letter.

She put the package on the table.b) ARGUMENT

contributed by the verb

She loaded the wagon with hay.

not a

profiled/obligatory participant

of verbc) ARGUMENT

contributed by construction

He baked her a cake.

She kicked him a ball.

She sneezed the foam off the cappuccino.

d) Traditional ADJUNCT

He baked a cake for her.

She broke the window with a hammer.

She swam in the summertime.

Figure 1: possible routes to argument status

The most common, prototypical case is one in which the profiled participant roles of the verb line up isomorphically with the argument roles of an argument structure construction. This is represented in cell (a) in the chart above. Another familiar case is one in which a non-profiled role is expressed by an adjunct construction as represented in cell (d).

In other cases, there is a mismatch between the verb’s and argument structure construction’s roles. Sometimes an argument role may not correspond to an independent obligatory participant role of the verb sense. For example, when the ditransitive construction is combined with verbs of creation, the recipient role is associated only with the construction; we do not need to assume that verbs of creation lexically specify a potential recipient. The same is true for certain verbs of motion as well. Kick for example only has two profiled participant roles; the recipient argument in She kicked him the ball is added by the construction.

The fourth logical possibility is that a profiled participant role of the verb is expressed by what is normally considered to be an adjunct phrase. As suggested in the (b) cell of the chart above, it seems appropriate to identify the with phrase that appears with load as an instance of this type. As discussed in the previous section, there are reasons to class the with phrase as a type of “intermediary” construction and other instances of the same construction (including what are usually referred to as instruments) normally function as adjuncts (in being omissible, able to appear sentence initially, after a clear adjunct such as yesterday, etc). However we have seen that the loaded-theme participant role of load is a profiled role. Because the with phrase codes a profiled role but is expressed by an phrase that is normally an adjunct, we might expect the behavior of this argument to fall somewhere in between that of traditional arguments and traditional adjuncts. In (55) we see that this is the case. While placing a clear adjunct before the with phrase is not crashingly bad in (55a), it is slightly less felicitous than the corresponding example in (55d). Other examples pattern similarly, depending on whether the participant coded by with corresponds to a profiled participant role of the verb or not:

(55) a. ?Pat loaded the wagon yesterday with hay.

b. ?Pat adorned the tree yesterday with lights.

c. Pat hit the wall yesterday with a stick.

d. Pat broke the window yesterday with a hammer.

5.3 Accounting for differences among instances of the same basic construction type

Rappaport and Levin (1985, 1988) have argued that the with phrase in Pat loaded the truck with hay is crucially distinct from instrumentals on the basis of the claim that certain related sentences receive different acceptability judgments. Several of their test frames can be seen to distinguish arguments from adjuncts. For example, consider (56) which invokes the classic do so test in which arguments are within the scope of do so VP anaphora, and adjuncts are outside it (cited judgments theirs):

(56)a. Liza covered the baby with a blanket and then Henry did so with a quilt.

b. *Liza loaded the wagon with hay and then Henry did so with straw. (Rappaport and Levin 1985)

Interestingly enough, cover was contrasted with load in (56a,b) by Rappaport and Levin on the assumption that the former licenses an instrumental while the latter licenses a distinct type of theme argument. However, in later work, Levin (1993) classifies cover as licensing the same construction as load the wagon with hay.

Of course there is a potential problem with (56b) which results from our world knowledge. It isn’t possible to load a wagon if it is already loaded. Notice (56b) is improved if we assume that the hay Liza loaded is removed before Henry puts straw onto the wagon. To the extent that there remains any difference in acceptability between (56a) and (56b), the difference may be attributed to the fact that the intermediary role corresponds to a profiled participant role of load but to an unprofiled participant role of cover. That is, cover, has three participant roles, the coverer, the cover, and the covered entity. The cover role is not profiled—it is not obligatory because its specific characteristics are typically not highly relevant. The goal of covering something is to keep that thing warm or to hide it. Exactly what is used to keep something warm or hide it is typically not essential to the discourse. Notice the cover role is easily omitted with an indefinite interpretation:

(57) She covered the baby. I wonder what she covered the baby with.

In this way, many differences can be attributed to the lexical semantics of the verbs involved; they do not necessarily necessitate treating the with phrases themselves as critically distinct.

To summarize, categorizing load with expressions as a type of causative + intermediary phrase, does not require that we be blind to any potential differences between uses with particular verbs. We need to account for verb meaning anyway, so it makes sense to look to verb meaning to determine whether differences in interpretation or in the range of possible paraphrases can be straightforwardly accounted for by it.

6. What is meant by surface form

In this section we clarify what is meant by surface form. Surface form need not specify a particular word order, nor even particular grammatical categories, although there are constructions that do specify these features. Adopting the notation of Goldberg (1992,1995) we might characterize the ditransitive construction as follows:

Sem: CAUSE-RECEIVE (agt rec theme)

| | |

verb ( )

Information-structure: rectopicality > theme topicality

Syn: Subj Obj1 Obj2

Figure 2: The Ditransitive Construction

The first line provides the semantics of the construction. The ditransitive involves a predicate with three arguments; these three arguments are labeled “agent” “recipient” and “theme” for convenience but there is no assumption that these thematic role labels are drawn from a universal or limited set. Instead the roles are determined by the meaning of the construction. In this case the main predicate is “CAUSE-RECEIVE” or more informally “give,” and the three argument roles correspond to the three major entities involved in the semantics of giving.

As is the case with other constructions, including words and morphemes, constructions typically allow for a range of closely related interpretations. The “CAUSE-RECEIVE” predicate associated with the ditransitive construction is subject to systematic variation depending on which verb class it interacts with. Thus the construction can be used to convey “intention to cause to receive” when used with verbs of creation; “refuse to cause to receive” when used with verbs of refusal, etc. (see Goldberg 1992, 1995; Kay Ms-2001, Leek 1996 for details and slightly differing analyses).

As indicated on the diagram in Figure 2 by the lines between the argument roles of the construction and the role array of the verb, the verb and its own arguments are integrated or fused with the predicate and arguments of the construction. Solid lines are used to indicate that the argument role of the construction must fuse with an independently existing participant role of the verb (recall cell a in Figure 1). Dashed lines are used to indicate that the argument role of the construction may be contributed by the construction without a corresponding role existing as part of the inherent verbal meaning. That is, a corresponding participant role may exist, but need not (recall cell c in Figure 1). The information structure row of information was not explicitly represented in earlier work, but its addition is straightforward. The specification noted is that the recipient argument should be more topical than the theme argument.

Finally, the linking of roles to grammatical relations is provided. See Goldberg (1995: chapter 4) for arguments that both generalizations and exceptional mappings can be captured by positing construction-specific linking generalizations when constructions are related within an inheritance hierarchy.[10]

From the representation above, it should be made clear that the reference to form in the definition abstracts away from specifics of surface form that can be attributed to other constructions. That is, an actual expression or construct typically involves the combination of at least half a dozen different constructions. For example, the construct in (57) involves the list of constructions given in (58a-g):

(57) What did Mina buy Mel?

(58) a. Ditransitive construction

b. Q-construction

c. Subject-Auxiliary inversion

d. VP construction

e. NP construction

f. Indefinite determiner construction

g. Mina, buy, Mel, what, do constructions

Constructions are combined freely to form actual expressions (constructs) as long as they can be construed as not being in conflict (invoking the notion of construal here is intended to allow for processes of coercion, see Michaelis (to appear)).

Thus, the same ditransitive construction is involved in active declarative form as well as in topicalized, clefted or questioned forms. That is, the recipient argument is an Object whether or not it appears directly after the verb or whether it appears as a distantly instantiated question word. It is, for example, the (non-echo) question construction that determines the fact that the wh-word appears sentence initially in English.[11]

7. Conclusion

The arguments in this paper should not be taken to imply that possible paraphrase relations play no role in the learning, processing or representation of language. The essentially structuralist observation that the semantic interpretation of one linguistic construct tends to be affected by the existence of possible alternatives, receives empirical support from a number of studies (e.g., Lambrecht 1994; Lambrecht and Polinsky 1997; Moore and Ackerman 1999; Spencer 2001).

In other work I have argued that the statistical use of paraphrases in actual discourse contexts is critical to unlocking Baker’s paradox of partial productivity (Goldberg 1993, 1995:122-125, see also Brooks and Tomasello 1999; Pinker 1984; Regier 1996). Paraphrase relations can also be seen to be relevant to on-line choices made in production (Bock, Loebell, and Morey 1992; Bock and Loebell 1990; Bock 1986).

However, it is less clear that one particular paraphrase should have a privileged status, nor that it is profitable to analyze one phrasal pattern solely by implicit or explicit reference to another. It has been argued here that by carefully examining a fuller range of surface phenomena, broader generalizations, surface generalizations in the form of Argument Structure Constructions, are revealed.

In accounting for similarities among alternative expressions and dissimilarities among instances of the same argument structure construction, careful attention must be given to the verb which is the same in the former and different in the latter. Recognizing surface generalizations surrounding argument structure (i.e., argument structure constructions) is important in that it leads to the recognition of generalizations in language that might otherwise be overlooked. But it is equally important to bear in mind that the meaning of a clause is more than the meaning of the Argument Structure Construction used to express it. Individual verbs as well as particular arguments and context must be factored in to the equation.

References

Alsina, Alex and Sam Mchombo. 1990. The syntax of applicatives in Chicheŵa: Problems for a Theta Theoretic Asymmetry. Natural Language and Lingusitic Theory 8 (4): 493-506.

Anderson, Stephen R. On the Role of Deep Structure in Semantic Interpretation. Foundations of Language 6 (1971): 197-219.

Arnold, Jennifer E., Thomas Wasow, Anthony Losongco, and Ryan Ginstrom. Heaviness vs Newness: the effects of structural complexity and discourse status on constituent ordering. Language 76, no. 1 (2000): 28-55.

Baker, Mark. Incorporation: A Theory of Grammatical Function Changing. Chicago: University of Chicago Press, 1988.

Bates, Elizabeth, Inge Bretherton, and Lynn Snyder. From First Words to Grammar: Individual Differences and Dissociable Mechanisms. New York: Cambridge University Press, 1988.

Bock, Kathryn J., Helga Loebe, and Randal Morey. From Conceptual Roles to Structural Relations: Bridging the Syntactic Cleft. Psychological Review 99, no. 1 (1992): 150-171.

Bock, Kathryn J., and Helga Loebell. Framing Sentences. Cognition V (1990): 1-39.

Bock, Kathryn J. Syntactic Persistence in Language Production. Cognitive Psychology 18 (1986): 355-387.

Bolinger, Dwight. Meaning and Form. London: Longman, 1977.

Bresnan, Joan. The Mental Representation of Grammatical Relations. Cambridge, Mass: MIT Press, 1982.

Bresnan, Joan. Locative Inversion and Universal Grammar. Language 70, no. 1 (1994): 72-131.

Brooks, Patricia, and Michael Tomasello. How children constrain their argument structure constructions. Language 75, no. 4 (1999): 720-738.

Brugman, Claudia. The Story of ‘Over’: Polysemy, Semantics and the Structure of the Lexicon Outstanding Dissertations in Linguistics. New York: Garland, 1988.

Bybee, Joan L. Morphology: A Study of the Relation between Meaning and Form: John Benjamins Publishing Company, 1985.

Bybee, Joan L. Regular Morphology and the Lexicon. Language and Cognitive Processes 10, no. 5 (1995): 425-455.

Chomsky, Noam. Remarks on Nominalizations. In Readings in English Transformational Grammar, ed. Roderick A. Jacobs and Pter S. Rosenbaum. Waltham, Massachusetts: Ginn and Company, 1970.

Culicover, Peter, and Ken Wexler. An application of the freezing principle to the dative in English: Social sciences Working Paper 39: UC Irvine.

Erteschik-Shir, N. Discourse Constraints on Dative Movement. In Syntax and Semantics, ed. Suzanne Laberge and Gillian Sankoff, 12, 441-467. New York: Academic Press, 1979.

Fillmore, Charles J. Indirect object constructions in English and the Ordering of Transformations. Janua Linguarum series.Mouton 1965.

Fillmore, Charles J. The Case for Case. In Universals in Linguistic Theory, ed. E. Bach and R.T.Harms, 1-88. New York: Holt, Rinehart and Winston, 1968.

Fillmore, Charles J. 1977. Topics in Lexical Semantics. In P.Cole (ed.) Current Issues in Linguistic Theory. 76-138. Bloomington: Indiana University Press.

Fillmore, Charles J. 1986. Pragmatically Controlled Zero Anaphora. In V. Nikiforidou, M. VanClay, M. Niepokuj and D. Feder (eds.) Berkeley Linguistic Society 12 92-77. Berkeley: Berkeley Linguistic Society.

Fillmore, Charles J., Paul Kay, Laura Michaelis and Ivan Sag. Construction Grammar. Stanford: CSLI, in progress.

Fraser, Bruce. A note on the Spray Paint Cases. Linguistic Inquiry 2 (1971): 604-607.

Garcia, Francisco Gonzalvez. Rethinking Predicational Be: From Copula Support to the Construction of a Lexical Grammar of Sentential Complementation in English. (2001-Ms).

Givón, Talmy. On Understanding Grammar. New York: Academic Press, 1979.

Gleitman, Lila, Gleitman Henry, Carol Miller, and Ruth Ostrin. Similar and Similar Concepts. : The Instute for Research in Cognitive Science, U of PA, 1995, IRCS Report.

Goldberg, Adele E. The inherent semantics of argument structure: The case of the English ditransitive construction. Cognitive Linguistics 3, no. 1 (1992): 37-74.

Goldberg, Adele E. 1993. Another Look at Some Learnability Paradoxes. Proceedings of the 25th Annual Stanford Child Language Research Forum . CSLI Publications.

Goldberg, Adele E. Constructions : a construction grammar approach to argument structure Cognitive theory of language and culture. Chicago: University of Chicago Press, 1995.

Goldberg, Adele E. The Emergence of Argument Structure Semantics. In B. MacWhinney (ed.) The Emergence of Language. Lawrence Erlbaum Publications, 1999.

Goldberg, Adele E. Patient arguments of causative verbs can be omitted: The role of information structure in argument distribution. Language Science 2000.

Goldberg, Adele E. Argument Realization: the role of constructions, lexical semantics and discourse factors. In Construction Grammar(s): Cognitive and Cross-language dimension, ed. Jan-Ola Östman: John Benjamins, to appear.

Green, Georgia M. Semantics and syntactic regularity. Bloomington,: Indiana University Press, 1974.

Gropen, Jess, Pinker Steven, Michelle Hollander, and Richard Goldberg. Syntax and Semantics in the Aquisition of Locative Verbs. Journal of Child Language 18, no. 1 (1991): 115-151.

Haiman, John. Iconicity in Syntax. Cambridge: Cambridge University Press, 1985.

Heine, Bernd. Grammaticalization Chains. Studies in Language 16 2, 1992. 335-368.

Hook, P.E. The English Abstrument and Rocking Case Relations. In Papers from the 19th Regional Meeting, Chicago Linguistics Society, 183-194. Chicago, 1983.

Hovav, Malka Rappaport, and Beth Levin. Building Verb Meanings. In The Projection of Arguments: Lexical and Compositional Factors, ed. Miriam Butt and Wilhelm Geuder, 97-134: CSLI Publications, 1998.

Iwata, Seizi. Locative Alternation and Two Levels of Verb Meaning. . Gifu University, Japan, 2000.

Iwata, Seizi. Does the wall really change its state? Toward the constructional meaning of the with variant in the locative alternation. Paper presented at the International Construction Grammar Conference in Helsinki, September 7, 2002.

Jackendoff, Ray. The Architecture of the Language Faculty. cambridge, Mass: MIT Press. 1997.

Jackendoff, Ray. 1996. The proper treatment of measuring out, telicity, and possibly even quantification in English. Natural Language and Linguistic Theory 14: 305-354.

Jackendoff, Ray. Foundations of Language. Oxford: Oxford University Press, 2002.

Jackendoff, Ray S. Semantic structures. Cambridge, Mass.: MIT Press, 1990.

Kay, Paul. Notes on argument structure constructions. manuscript (1997).

Kay, Paul. Argument Structure Constructions and the Argument-Adjunct Distinction. Ms-2001.

Keenan, Edward L. 1976. Towards a Universal Definition of “Subject” In C.N. Li (ed.) subject and topic, 303-334. New York: Academic Press.

Keenan, Edward L. 1984. Semantic Correlates of the Ergative/Absolutive Distinction. Linguistics 22: 197-223.

Lakoff, George. Women, fire, and dangerous things : what categories reveal about the mind. Chicago: University of Chicago Press, 1987.

Lambrecht, Knud. Information Structure and Sentence Form. Cambridge: Cambridge University Press, 1994.

Lambrecht, Knud, and Maria Polinsky. Typological Variation and Sentence-Focus Constructions. CLS 33, 1997.

Langacker, Ronald W. Foundations of cognitive grammar Volume I. Stanford, Calif.: Stanford University Press, 1987a.

Langacker, Ronald W. Nouns and Verbs. Language 63 1987b: 53-94.

Langacker, Ronald W. Foundations of Cognitive Grammar Volume II. Stanford: Stanford University Press, 1991.

Larson, Richard. On the Double Object Construction. Linguistic Inquiry 19 (1988): 335-392.

Leek, Frederike Van Der. Rigid Syntax and Flexible Meaning: The Case of the English Ditransitive. In Conceptual Structure, Discourse and Language, ed. Adele E. Goldberg, 321-346. Stanford: CSLI Publications, 1996.

Lees, Robert B. The grammar of English nominalizations. The Hague: Mouton, 1960.

Levin, Beth. English Verb Classes and Alternations. Chicago: Chicago University Press, 1993.

Lindner, Susan. A Lexico-semantic Analysis of Verb-Particle Constructions with Up and Out. dissertation, UCSD, 1981.

Meyer, D. E., and R. W. Schvaneveldt. Facilitation in recognizing pairs of words: Evidence of a dependence between retrieval operations. Journal of Experimental Psychology 90, no. 2 1971: 227-234.

Michaelis, Laura. Implicit and Explicit Type-Shifting in Construction Grammar. Cognitive Linguistics to appear.

Michaelis and Ruppenhofer. Beyond Alternations: a constructional model of the German applicative Pattern. Stanford: CSLI Publications.2001.

Moore, John, and Farrell Ackerman. Syntagmatic and Paradigmatic Dimensions of Causee Encoding. Linguistics and Philosophy 22, 1999: 1-44.

Nakajima, Heizo. Koobun bunpoo no kihon o kangaeru (reflections on the basic assumptions of construction grammar) The Rising Generation CXLVII, no. 1 (2002): 34-37. Translated for me by Tsuguro Nakamura.

Newmeyer, Frederick. Theoretical Implications of Phrasal Category–Grammatical Relations Mismatches. In Linguistic Mismatches, ed. Laura Michaelis and Elaine Francis. Stanford: CSLI Publications, to appear.

Oehrle, Richard. Discourse Constraints on Dative Movement. In Syntax and Semantcs, 12, 1975.

Partee, Barbara Hall. Subject and Object in Modern English Dissertation. Published in Ou

“构式语法”重要文献之一

Raymond W. Gibbs, Jr

Introduction

Embodiment in the field of cognitive science refers to understanding the role of an agent’s own body in its everyday, situated cognition. For example, how do our bodies influence the ways we think and speak? Consider the following narrative written by a 23-year-old woman, Sandra, who was asked to describe a recent, important life event. Sandra began her narrative by noting that she was engaged to be married to an older man who worked in the computer industry in northern California. Quite recently, Sandra’s fiancé asked her to sign a prenuptial agreement and this request evoked many feelings that Sandra struggled to deal with.

I know that I shouldn’t be so naive about this sort of thing, but when he presented me with a draft of the agreement, it was so formal and legal and felt so cold to me that I just broke down crying. I simply couldn’t stand to see our future relationship be reduced to questions of money. It seemed like Barry didn’t trust me, or that he lacked faith in our future. I had always thought that we were in this together, going forward as partners as we started dating, got serious, then engaged, and hopefully soon married. Now my parents want me to consult with a lawyer to insure I don’t get screwed by the pre-nup.

I’m trying hard to find the right balance between understanding Barry’s needs to protect himself and my own needs for emotional security…I’m trying to be flexible about the whole thing…I love Barry and I know he loves me and I wish that the feeling of love would be enough to sustain us through anything. But the idea of getting divorced, even before we have been married, makes me ill. Everyone tells me that I’ll get over this and that doing the pre-nup is probably the right thing to do. That may be so. The wedding is in August. Hopefully by that time, I’ll be mellow about what we’re going through right now.

This narrative is not particularly remarkable in terms of how Sandra described her recent experience. However, a closer look at what is said reveals how various embodied experiences help structure the narrative. For example, Sandra commented early on that “I couldn’t stand to see our future relationship be reduced to questions of money,” referring to the physical experience of standing, or failing to remain standing, to describe how she felt about her relationship becoming so centered on money issues. Later on, Sandra said, “I had always thought that we were in this together, going forward as partners as we started dating, got serious, then engaged, and hopefully soon married.” At this moment, Sandra clearly talked of her relationship in terms of being physically together with her boyfriend as they started out on a journey, beginning when they first began dating, soon traveling to the point of getting serious, and then moving forward along a path toward the eventual destination of marriage.

Sandra also noted her struggle “to find the right balance between understanding Barry’s needs to protect himself and my own needs for emotional security.” This emotional experience is referred to metaphorically, as if Sandra were physically balancing two opposing weights while trying to remain upright. As she worked to come to terms with her fiancé’s request for a prenuptial agreement, Sandra was “trying to be flexible about the whole thing,” again showing that she is conceptualizing her emotional experience as if her body must adjust to remain flexible in order not to be injured when confronted with physical burdens. Finally, Sandra hoped for her wedding that “by that time, I’ll be mellow about what we’re going through right now,” referring to the physical obstacle that she and her fiancé were struggling to overcome along the path of their relationship journey.

Sandra’s narrative illustrates how the ways we think about our experiences may be shaped by embodiment. She specifically talked of her mental/emotional experiences in terms of recurring patterns of embodied action (e.g., standing, being flexible, movement along paths toward goals, remaining balanced). Sandra was likely not conscious of the embodied character of her words, and readers probably do not think of her emotional experiences as specific embodied actions. Yet Sandra’s description of her emotional experiences in terms of embodied action is not a linguistic accident, but demonstrates how embodiment provides the foundation for how people interpret their lives and the world around them.

What must a body be like for it to support cognition, language, and consciousness? Did Sandra’s embodied experience shape the way she thought about particular topics, or did she merely talk that way? One of the traditional beliefs in the cognitive sciences is that intelligent behavior, including the ability to perceive, think, and use language, need not arise from any specific bodily form. Thermostats, computers, robots, and brains in vats may all, under the right circumstances, exhibit sophisticated cognitive skills. Under this view, cognitive systems are best characterized in terms of their functional states (i.e., their logical and computational processes) without concern for how these states are physically realized (i.e., as human brains, silicon chips, or robots). The building materials that shape the contents of mental life simply do not matter. Minds may be realized in flesh, silicon, or even cream cheese (Putnam, 1975). To be in a specific mental state is simply to be in a physical device of whatever type satisfies a specific formal/functional description.

This traditional conception of mind and body has imposed serious limitations on the scholarly study of mental life in cognitive science. Although psychologists and others readily admit that much knowledge is derived from sensory perception, few scholars, until recently, have emphasized the importance of kinesthetic action in theoretical accounts of how people perceive, learn, think, experience emotions and consciousness, and use language. This book advances the idea that the traditional disembodied view of mind is mistaken, because human cognition is fundamentally shaped by embodied experience. My aim is to describe the way in which many aspects of cognition are grounded in embodiment, especially in terms of the phenomenological experience of our bodies in action. Embodiment may not provide the single foundation for all thought and language, but it is an essential part of the perceptual and cognitive processes by which we make sense of our experiences in the world.

Why has cognitive science been so neglectful of embodiment in constructing theories of perception, cognition, and language? The denial of the body in consideration of human thought has been part of the Western intellectual tradition since the time of the ancient Greeks. Perhaps the best voice for this earlier view was Plato, as shown in the following dialogue from the “Phaedo”:

All these considerations, said Socrates, must surely prompt serious philosophers to review the position in some such way as this…So long as we keep the body and our soul contaminated with this imperfection, there is no chance of our ever attaining satisfactorily to our object, which we assert to be the truth….The body fills us with loves and desires and fears and all sorts of nonsense, with the result that we literally never get an opportunity to think at all about anything…That is why, on all these accounts, we have so little time for philosophy….It seems, to judge from the argument, that the wisdom which we desire and upon which we profess to have set our hearts will be attainable only when we are dead…It seems that so long as we are alive, we shall continue closest to knowledge if we avoid as much as we can all contact and association with the body, except when they are absolutely necessary, and instead of allowing ourselves to become infected with its nature, purify ourselves from it until God himself gives us deliverance. (Hamilton & Cairns, 1961: 49)

Plato viewed the body as a source of distraction in intellectual life that must be eradicated in the practice of philosophy. Separation of the mind and body and the hierarchical ordering of mind over body haunt the history of Western philosophical accounts of knowledge from Plato, Aristotle, and Augustine through to Descartes and Kant. For example, in early Christian writings, bodily sensations and desires were rivaled in contests against a higher form of Truth, or closeness to God. As St. Augustine wrote in the fifth century, “More and more, O Lord, you will increase your gift in me, so that my soul may follow me to you, freed from the concupiscence which binds it, and rebel no more against itself” (Augustine, 1961: 234). Augustine fixed the body as a source of sin, weakness, and the measure against which the strength of his will toward God is knowable.

In the 17th century, Rene Descartes’ struggle with a purely material body and a perfectly insubstantial mind led him to propose that the body is, in fact, an idea in the mind (Descartes, 1984, 1985). The body’s materiality, along with other objects that are impressed upon body substances, is a literalization of this idea in our experience. When we pay attention to it, the body materializes, and we become aware of the body as an object. However, as our attention centers on other things, or on thought itself, the body disappears.

Mental phenomena, according to Descartes, have no place in the quantifiable world of physics, but have a completely autonomous status: “I am a substance the whole nature or essence of which is to think, and which for its existence does not need any place or depend on any material thing” (Descartes, Discourse, Part IV). Descartes distinguished, then, between physical substances (“res extensa”), which can be measured and divided, and thinking substances (“res cogitans”), which are unextended and indivisible. The human body, including the brain and nervous system, belongs to the first group, whereas the mind, including all thoughts, desires, and volitions, belongs to the second.

Cartesian dualism arose from Descartes’ claim that he could doubt the existence of physical objects, including his own body, but not the existence of his thoughts or thinking. Although Descartes worried about possible interactions of mind and body, Cartesian dualism evolved into an epistemological tradition that separated the mind as rational, thinking, immaterial, and private from the body as an irrational, corrupt, and physical substance that merely provided public, physical exertion on the material world. This bifurcation of the person into mind and body has subsequently given rise to many other dualisms, including subjective as opposed to objective, knowledge as opposed to experience, reason as opposed to feeling, theory as opposed to practice, and verbal as opposed to nonverbal. Cartesianism has also led to the romantic view of the body as the last bastion of what is natural, unspoiled, preconceptual, and primitive in experience. Bodily movement is viewed as behavior, with little relevance to language, thought, or consciousness, and not as meaningful action.

The Western tradition since Descartes has generally assumed that the body is a solid object and the self, in particular the mind, is an ethereal subject mysteriously infused into the body object. Throughout history, the mind has been modeled as a series of different material objects (e.g., a hydraulic machine, a telephone switchboard, a hologram, a digital computer). Cognitive science, as an interdisciplinary research enterprise, came into being in the 1950s with the rise of the MIND IS A COMPUTER metaphor, which resulted from technological advances in computing machinery. Alan Turing (1950) outlined a method for assessing the question “Can machines think?” Following in Descartes’ footsteps, Turing emphasized the importance of drawing a “fairly sharp line” (p. 434) between a person’s physical capacities and his/her intellectual abilities. Turing asked us to consider a scenario that included three people – a man (A), a woman (B), and an interrogator of unspecified sex (C). The interrogator was in a separate room from the man and the woman, and the interrogator’s task was to determine which of the two was a man and which was a woman on the basis of their written answers to certain questions (e.g., “What is the length of your hair?”). It is A’s task to confuse the interrogator and B’s task to help. The test proper comes into play by swapping the man (A) with a machine. If the interrogator makes the same set of judgments, deductions, and guesses after the swap as before (i.e., the interrogator is unable to distinguish the machine’s answers from the man’s answers), then the machine has passed the “Turing test.” The machine whose behavior is indistinguishable from the intellect of the man is the machine that thinks.

Cognitive science models of intelligent human activity have mostly continued to assume, like Turing, that cognition is autonomous, logical, and disembodied. In his history of cognitive science, Gardner (1985) claimed that the exclusion of the body was, in fact, a benign methodological decision: “Though mainstream cognitive scientists do not necessarily bear any animus against the affective realm, against the context that surrounds any actor or thought, or against historical or cultural analyses, in practice they attempt to factor out these elements to the maximum extent possible…This may be a question of practicality: if one were to take into account these individualizing and phenomenalistic elements, cognitive science might well become impossible” (p. 41).

Some cognitive scientists question whether the exclusion of the phenomenological body, along with other aspects of experience such as emotion and consciousness, is merely a methodological issue, and not really constitutive of what cognitive scientists believe is essential about cognition. Of course, many scholars now try to avoid the strict separation of mind and body assumed by Cartesian dualism. The most popular strategy, especially in recent decades, has been to reduce mental events to brain processes and replace internal explanations with instrumental ones. In some cases, the reduction of mind to brain carries with it the reduction of body to brain. Neuroscientists, for instance, seldom acknowledge the role played by the body as a whole in the cognitive operation of the brain. The body is reduced to its representation in the somatosensory cortex and is considered important only to the extent that it provides the raw sensory input required for cognitive computations. In other cases, the body is first reduced to the mind, and then reduced to the brain. This is especially true in psychology, where the body is first treated as an intentional object (i.e., an image, a mental representation) and then reduced to neural computations.

Contemporary philosophers argue over whether a physical body is necessary for knowledge and cognition, often by considering the implications of different thought experiments in which the mind may be divorced from critical aspects of bodily experience. For instance, consider the following scenario:

Imagine a brilliant neuroscientist named Mary, who has lived her entire life in a room that is rigorously controlled to display only various shades of black, white, and grey. She learns about the outside world by means of a black/white television monitor, and being brilliant, she manages to transcend these obstacles. She becomes the world’s greatest neuroscientist, all from within this room. In particular, she comes to know everything there is to know about the physical structure and activity of the brain and its visual system, of its actual and possible states. (Churchland, 1985: 22)

Philosophers argue, based on examples like the above, over whether qualia (i.e., the phenomenal character of our experience), such as one’s subjective sensations of color, must be mental states that are causally related to the neurophysiology of the brain (see Churchland, 1984; Jackson, 1982, 1986). These scenarios, however, dramatically fail to recognize the need for a real living body in knowing about the world. There is no acknowledgment of Mary as a living person, made of flesh, blood, and bone, who moves and has awareness of the felt qualities of her own actions. Mary’s first-person experiences of her own body in relation to the environment provide knowledge that is “qualitatively incommensurate” with whatever may be happening in her own brain, or anyone else’s (Sheets-Johnstone, 1999: 167). Mary learns about qualia because she subjectively experiences them through her own bodily actions. Knowledge of a set of abstract propositions, such as Mary’s understanding of the neurophysiology of color vision, means nothing unless a person experiences in some embodied sense the physical world to which these propositions refer (Sheets-Johnstone, 1999).

Cognitive psychologists, like many philosophers, often fail to recognize the significance of embodied action in the study of human mental life. Most experimental investigations of perception and cognition occur in laboratory situations where a person passively observes stimuli and then responds in some specified manner to what has been presented. In some instances, the person is physically restricted in his or her movements (e.g., head rests are used in psychophysical experiments). In cases where the participant must move to respond to stimuli, such as having to push a button or speak aloud, psychologists work hard to eliminate the movement from their theoretical understanding of the processes involved in perception and cognition. Cognitive processes, especially, are viewed as strictly mental phenomena that have little to do with embodied experience. The body is the vessel for the mind and brain, but has negligible importance in characterizing the essence of mental life.

However, the situation is now changing. Consider just three examples of how psychologists now pay attention to embodied action when studying different cognitive phenomena. First, the classic empirical work on mental imagery investigates possible correspondences between mental imagery and visual perception. For example, participants in one classic study were presented with two-dimensional drawings of pairs of three dimensional objects. The participants’ task was to determine whether the two represented objects were identical except for orientation (Shepard & Metzler, 1971). Some of the figures required rotation solely within the picture plane, whereas others required rotation in depth (“into” the page). The general result was that, whether for two- or three-dimensional rotations, participants seemed to rotate the objects mentally at a fixed rate of approximately 60 degrees/second. For many years, psychologists assumed that cognitive abilities, such as those observed in mental rotation studies, demonstrate the tight link between visual perception and mental imagery. Although numerous studies examine people’s kinesthetic and motor imagery, scholars traditionally have not searched for explicit relations between kinesthetic activity and mental imagery.

However, recent work suggests that many aspects of visual and motor imagery share a common representational, and possibly neuropsychological, substrate. Various studies demonstrate that the ability to transform mental images is linked to motor processes, so that rotating one’s hands in the direction opposite to the required mental rotation slows down the speed of mental rotation (Wexler, Kosslyn, & Berthoz, 1998). Researchers now claim that “visuomotor anticipation is the engine that drives mental rotation” (Wexler et al., 1998). Under this view, similar mechanisms drive both visual image transformation and the production of embodied movements. The ability to plan movements as simulated actions, and not as actual motor plans, may be the common element underlying embodied action and mental imagery performance (Johnson, 2000). These new developments in cognitive psychology illustrate how correcting for a previous neglect of embodied experience in experimental studies leads to a richer picture of the importance of embodiment in human cognition.

Psycholinguists have also slowly begun to seek out the embodied foundation of linguistic structure and meaning. Recall Sandra’s earlier comment in response to her fiancé’s request for a prenuptial agreement that “I couldn’t stand to see our future relationship be reduced to questions of money.” Why is it that Sandra used the word “stand” to refer to an abstract, mental experience of her adjusting to her fiancé’s demand? Traditional studies on how people process ambiguous, or polysemous, words such as “stand” generally assume that each sense of a word is listed as part of its entry in the mental lexicon. For example, do people immediately access all the possible senses for the word “stand,” with context determining which meaning is appropriate afterward? Or does context constrain lexical access so that only the correct meaning of “stand” is accessed during immediate utterance interpretation? These empirical questions have been studied extensively (Gorfein, 2001).

Psycholinguists rarely ask whether people have intuitions about why “stand,” or any polysemous word, has the variety of meanings it does. Recent studies, however, demonstrate that people’s intuitions about the meanings of “stand” are shaped by their embodied experiences of standing (Gibbs, Beitel, Harrington, & Sanders, 1994). Thus, people tacitly recognize that Sandra’s use of “stand” has a metaphorical meaning that is related to their embodied experiences of struggling to remain physically upright when some physical force acts against them. People’s understandings of linguistic meanings are not divorced from their embodied experiences, but rather are fundamentally constrained by them in predictable ways.

Following Piaget’s early writings, developmental psychology has also started to meaningfully explore how embodied action may underlie children’s acquisition of perceptual/conceptual knowledge. For example, infants’ interest in things that move assists them in understanding some cause-effect relations in the physical world. Sophisticated studies indicated that infants 12 months old and younger are capable, in the right setting, of making causal attributions to the behavior of objects they see in the world (Gergely, Nadasdy, Csiba, & Biro, 1995; Spelke, Philip, & Woodward, 1995). The infant’s developing sensitivity to causal relations may underlie the acquisition of a concept for agency (i.e., things move because of internal forces or human intentions).

These studies, however, despite their brilliance, situate the child as a passive observer who learns to reason about the physical world by visual inspection of real-world events. Several experiments now demonstrate the importance of the child’s bodily exploration of the physical world in learning about objects and their behaviors (Adolph, 1997, 2000; Bertenthal, Campos, & Kermoian, 1994; Hertenstein, 2002; Needham, Barrett, & Peterman, 2002). This empirical work suggests that many basic concepts may arise from rudimentary bodily actions and young children’s felt experiences of them. Causation and agency, for example, may be rooted in infants’ phenomenological sense of their own bodies’ interactions with objects and other people. Even before infants possess any ability to physically manipulate objects with their hands and feet, they directly experience cause and effect from the movement of their lips, tongues, and mouths during breastfeeding, or from chewing food, which transforms it to something that can be swallowed easily. An encouraging trend in developmental psychology is greater attention given to infants’ phenomenological experience in relation to cognitive growth.

These brief examples illustrate how looking for embodied action in thought and language may provide a different picture of human cognition than has traditionally been assumed within cognitive science. Much recent work in cognitive science views embodiment as a matter of brain states and neural activity. We have indeed learned a great deal from these neuroscientific studies. However, as Roger Sperry noted over sixty-five years ago, “An objective psychologist, hoping to get at the physiological side of behavior, is apt to plunge immediately into neurology trying to correlate brain activity with modes of experience. The result in many cases only accentuates the gap between the total experience as studied by the psychologist and neuronal activity as analyzed by the neurologists. But the experience of the organism is integrated, organized, and has its meaning in terms of coordinated movement” (1939: 295).

The psychologist Scott Kelso more recently suggested, “It is important to keep in mind…that the brain did not evolve merely to register representations of the world; rather, it evolved for adaptive actions and behaviors. Musculoskeletal structures coevolved with appropriate brain structures so that the entire unit must function together in an adaptive fashion…it is the entire system of muscles, joints, and proprioceptive and kinesthetic functions and appropriate parts of the brain that evolve and function together in a unitary way” (1995: 268).

The brain is certainly part of an integrated dynamic system devoted to the moment-by-moment embodied dynamics of everyday life. Viewing the brain simply as an information-processing or computational device, as the center of cognition, ignores the centrality of animate form in human thought (Sheets-Johnstone, 1999).

This book describes the ways that perception, concepts, mental imagery, memory, reasoning, cognitive development, language, emotion, and consciousness have, to varying extents, groundings in embodiment. My strategy in exploring the significance of embodiment in the study of these topics adopts what may be called the “embodiment premise”:

People’s subjective, felt experiences of their bodies in action provide part of the fundamental grounding for language and thought. Cognition is what occurs when the body engages the physical, cultural world and must be studied in terms of the dynamical interactions between people and the environment. Human language and thought emerge from recurring patterns of embodied activity that constrain ongoing intelligent behavior. We must not assume cognition to be purely internal, symbolic, computational, and disembodied, but seek out the gross and detailed ways that language and thought are inextricably shaped by embodied action.

The key feature of this premise is the idea that understanding the embodied nature of human cognition demands that researchers specifically look for possible mind-body and language-body connections. Understanding embodied experience is not simply a matter of physiology or kinesiology (i.e., the body as object), but demands recognition of how people dynamically move in the physical/cultural world (i.e., the body experienced from a first-person, phenomenological perspective). The mind (its images, thoughts, representations) is created from ideas that are closely related to brain representations of the body and to the body’s continued activities in the real world.

Fortunately, there is an accumulating body of empirical evidence showing how embodied activities shape human cognition. In the spirit of cognitive science, this “empirical” evidence includes data collected from controlled laboratory studies, naturalistic field observations, neuropsychological case studies, linguistic research, artificial intelligence (and artificial life) modeling, and various phenomenological studies and reports. To be sure, many of the scholars whose studies are described here may not entirely agree with my interpretation of their work as support for “embodied” cognition. Some of these disagreements center around what is meant by the terms “embodied” and “embodiment.” I argue that “embodiment” may refer to, at least, three levels of personhood (see Lakoff & Johnson, 1999): neural events, the cognitive unconscious, and phenomenological experience. Although amazing advances have been made in understanding neural processes, insignificant attention has been given to people’s phenomenological experience in explaining many aspects of perception, cognition, and language. I address this problem in the pages that follow.

At the same time, special emphasis will be given in the following chapters to two important developments in cognitive science. The first is the approach to cognition known as dynamical systems theory. Dynamical approaches emphasize the temporal dimensions of cognition and the ways in which an individual’s behavior emerges from interactions of brain, body, and environment. Simple and complex behavioral patterns are higher-order products of self-organization processes. Virtually all living organisms self-assemble, or are self-organizing systems, “as emergent consequences of nonlinear interaction among active components” (Kelso, 1995: 67). Self-organized patterns of behavior emerge as stable states from the interaction of many subsystems. Yet the emerging higher-order behavior is also capable of “enslaving” lower-level components in such a way that behavioral patterns can often be described by relatively few dimensions. Much of the emphasis, then, in dynamical systems theory is on the structure of spaces of possible behavioral trajectories and the internal and external forces (i.e., couplings between brain, body, and world) that shape these trajectories as they unfold. More specifically, dynamical systems theory is a set of mathematical tools that can be applied to characterize different states of the system as these evolve in time. In this way, a dynamical view aims to describe how the body’s continuous interactions with the world provide for coordinated patterns of adaptive behavior, rather than focusing on how the external world become represented in the inner mind.

Francisco Jos?Ruiz de Mendoza Ibáñez

Olga Isabel Díez Velasco

University of La Rioja

0. Introduction

Ever since George Lakoff and Mark Johnson (1980) published their seminal work Metaphors We Live By, many cognitive linguists have devoted much of their research to exploring metaphorical systems in different languages. Metonymy, however, has received comparatively little attention. During the eighties and the nineties work on metonymy has mainly focused on setting up definitional and typological criteria (Croft, 1993; Dirven, 1993; Langacker, 1993; Kövecses and Radden, 1998, 1999), studying the metonymic grounding of metaphor (Barcelona, 2000; Radden, 2000), its role in conceptual interaction (Goossens, 1995; Ruiz de Mendoza, 1997a; Díez, 2000; Turner and Fauconnier, 2000) and in inferencing (Gibbs, 1994; Thornburg and Panther, 1997; Panther and Thornburg, 1998, 1999; Ruiz de Mendoza, 1999a; Pérez and Ruiz de Mendoza, 2001). Most of this research, however, has been concerned with the conceptual level of analysis, whereas the study of the potential impact of metonymy on linguistic structure has been largely neglected[2]. In the present paper we intend to show that metonymy does have that kind of impact. In order to do so, we start by giving an outline of what we believe is a reliable definition of metonymy as a cognitive mechanism. Then, we determine two criteria for a classification of metonymic types: (i) the degree of genericity and (ii) the ontological nature of the domains involved. The first criterion allows us to distinguish between high and low level metonymies; the second, between situational and non-situational metonymies. In this connection, we observe that only high-level metonymies are relevant from the point of view of linguistic structure. We finally note that while non-situational high-level metonymies relate to clause-internal processes, situational high-level metonymies seem to account for much of the conventional value of some illocutionary constructions[3].

1. Defining metonymy

In Cognitive Linguistics metonymy has generally been described as one form of what Lakoff (1987) calls an idealised cognitive model (or ICM). An ICM is an organised conceptual structure i.e. a knowledge domain which results from the activity of a structuring principle. Metonymy was first described in cognitive terms by Lakoff and Johnson (1980: 39) as a process which “allows us to conceptualize one thing by means of its relation to something else? However, this definition is not sufficient to understand the nature of this phenomenon and its relation to metaphor. In a more refined account, Lakoff and Turner (1989) have described metonymy as a conceptual mapping within a single domain which involves a ‘stand-for?relationship and has mainly a referential function.

However, it may be observed that Lakoff and Turner’s (1989) definition of metonymy is weak in that it does not provide truly definitional features but only some usual characteristics. Thus, metonymy may be used non-referentially, as in Mary is just a pretty face (‘Mary’s only worth is having a beautiful face? and both the ability to refer and the resulting ‘stand-for?relationship is also possible in the case of some metaphors, as in There is the pig waiting!, where by “pig?is meant ‘the person who behaves like a pig?(i.e. ‘whose behaviour is oppressive and nasty just as we may think a pig’s behaviour is? cf. the more common predicative use of “pig?in John is a pig ‘John is oppressive and nasty?. In order to make up for these weaknesses, we contend that metonymy is best described according to two parameters: (a) the nature of the relationship between the source and target domains, and (b) the kind of mapping process involved. Regarding the former, we argue that metonymy is always based on a domain-subdomain relationship, where the main domain is referred to as the matrix domain, and that only two basic types of metonymy can be distinguished: one in which the source is a subdomain of the target, as in The sax won’t come today, where ‘the sax?is a subdomain of ‘the sax player? and another in which the target is a subdomain of the source, as in He always enjoys Shakespeare, where by “Shakespeare?is meant ‘his literary work? which is a subdomain of our knowledge about the famous playwright (cf. Ruiz de Mendoza, 1997b, 2000). Metonymies in the first group are labelled source-in-target metonymies and those in the second group target-in-source metonymies. This distinction, which rejects the existence of part-for-part metonymies, proves relevant to explain a number of linguistic and communicative phenomena. For example, when a source-in-target metonymy provides the antecedent for an anaphoric pronoun, reference is consistently made to the target domain of the metonymy (e.g. The sax won’t come today; he/*it has the flu), which is the main domain of reference or matrix domain; in cases of target-in-source metonymies, on the other hand, it is the source domain, which is also the matrix domain, that is selected for anaphoric reference (e.g. She loves Shakespeare; she reads him/*it a lot). In general, source-in-target metonymies work on the basis of domain expansion (i.e. we are provided with global access to the full matrix domain just by invoking one of its subdomains), while target-in-source metonymies make use of domain reduction (i.e. only a subdomain of the matrix is relevant). Note that target-in-source metonymies make use of a well-defined conceptual domain (i.e. the matrix domain) to refer to a subdomain which is not necessarily as clearly identifiable. As a result, this kind of metonymy turns out to be an excellent communicative resource whenever the speaker finds himself incapable of either expressing the intended referent or of determining its exact nature. For example, in order to interpret correctly the metonymy in Chrysler has laid off a hundred workers, it is not necessary for the hearer to pin down with accuracy the actual referent of the matrix domain “Chrysler?(it could be the head of a department, or someone under him, or a whole board, or a special committee in charge of employment regulations, etc.). We also understand, by invoking the matrix domain in this way, that it is the whole company, rather than just one of its employees, that is responsible for firing the workers. This is a consequence of the special nature of this metonymic shift where the matrix domain in being, as noted above, a well-defined conceptual domain, figures more prominently in the hearer’s mind than any of its subdomains. In contrast, in source-in-target metonymies both the source and the target are well-defined domains, which prevents a similar kind of communicative effect from taking place.

As far as the mapping process is concerned, we postulate that metaphors, which are mappings or sets of correspondences across discrete domains, can belong to two main types: many-correspondence and one-correspondence metaphors (cf. Ruiz de Mendoza, 1997b). Many-correspondence metaphors have a fully-fledged system of correlations where the structure of the target is used to reason about the source. For example, in LOVE IS A JOURNEY travellers map onto lovers, travellers?common destinations onto lovers?common goals, the vehicle onto the love relationship, impediments to travel onto difficulties in the relationship, and so on. The situation is different in the case of one-correspondence metaphors, where the fact that there is only one relevant correlation between source and target relegates structural relationships to a secondary role. Instead, in metaphors of this kind, the mapping has the function of singling out a specific feature of the source which is then to be attributed to the target. A good example is the metaphor PEOPLE ARE ANIMALS, which helps us to understand various forms of human behaviour, skill, and (subsidiarily) appearance in terms of corresponding animal behaviour, skill, and appearance as culturally attributed or as observed. Thus, in this metaphor, a pig is nasty and oppresive, a lion is courageous, a rat is disloyal or deceitful, an eagle is good at noticing small details (e.g. Mary has an eagle eye), a bull is clumsy (because of its size), and so on.

Interestingly enough, metonymies are always one-correspondence mappings. For example, in Bush attacked Iraq, there exists only one correspondence between ‘Bush?(source) and ‘the U.S. army? (target). Notice should be taken that metonymies are closer to one-correspondence than to many-correspondence metaphors. The number of correspondences and the nature of the domains are connected to the referential or predicative use of metaphor and metonymy. Only one-correspondence mappings can be used referentially. Metonymies, as we have already pointed out, can be used both referentially (cf. Bush attacked Iraq) and predicatively (cf, Mary is just a pretty face). Likewise, one-correspondence metaphors can have both a referential (e.g. There’s the pig waiting) and a predicative function (e.g. John is a pig). However, while the predicative potential of metonymies is low, their referential potential is high. Conversely, one-correspondence metaphors have a high predicative potential and a low referential potential. The low predicative potential of metonymies is probably due to the fact that it is virtually impossible to map structural relationships in the case of domain/subdomain mappings since the domain of reference (i.e. the matrix domain) for a given subdomain is also the source or the target of the metonymic mapping itself. As a result, in order to have a predicative use of a metonymy either the context of situation or the linguistic context must act as cues for the hearer to determine the nature of the quintessential characteristic to be mapped. In this respect, predicative metonymies resemble one-correspondence metaphors to a considerable extent. Thus, in Mary is just a pretty face, one singular feature of ‘face?which is parametrized by the adjective “pretty? is singled out and ascribed to Mary. However, since Mary is the possessor of the face that is being referred to, the relationship between the source and target is domain-internal and, therefore, metonymic. On the other hand, one-correspondence metaphors can be used referentially because of the structural relationship which holds between the feature which is highlighted for interpretation and the rest of the domain. In this process we are provided with access to an extremely rich description which is the basis for the referential potential of the metaphorical expression. Since it is not just an isolated feature but a whole conceptual domain that is exploited in referential metaphor, the situation is no different from what is the case in referential metonymy, where the whole matrix domain is involved.

Our discussion thus points to the existence of a continuum between metaphor and metonymy with many-correspondence metaphors and referential uses of metonymy occupying the end points, as the following figure illustrates:

One-correspondence One-correspondence

metaphor metaphor

(predicative uses) (referential uses)

* * * * *

Many-correspondence Metonymy Metonymy

metaphor (predicative uses) (referential uses)

Figure 1: Metaphor-metonymy continuum

2. Levels of metonymic description

Developing taxonomies of metonymies is one of the crucial concerns of current research in Cognitive Linguistics (Dirven, 1993; Kövecses and Radden, 1998, 1999; Panther and Thornburg, 1999). The motivation underlying classifications is in most cases connected to the desire to understand the exact nature of metonymy. One particularly appealing proposal has been offered by Panther and Thornburg (1999) who have classified metonymies according to the scope of action of the metonymy and have distinguished three main groups: referential, predicative and illocutionary. In referential metonymies a single concept stands for another (e.g. She’s been dating a farm hand, where ‘hand?stands for ‘worker?; in predicative metonymies a statement is used to refer to a different statement (e.g. I’ll be brief stands for ‘I will speak briefly?; and in illocutionary metonymies one illocutionary type stands for another (e.g. the assertion I don’t know where the bus stop is stands for the question Where is the bus stop?). The significance of Panther and Thornburg’s classification lies in the fact that for them metonymy is not restricted to the referential type and research on metonymy needs to be carried out from a broader perspective. Their typology, however, does not explain why some metonymies have an impact on grammatical organization while others do not. Besides, Panther and Thornburg’s account does not encompass all kinds of metonymy as evidenced in (1):

(1) The poor dog left with its tail between its legs.

Imagine that sentence (1) is uttered in a situation in which a man is talking to a friend about the punishment his dog has received. Here, we find a metonymic mapping in which the actual leaving of the dog with its tail between its legs stands for the whole situation in which a dog after being punished runs away in that manner[4]. Since this metonymy does not square well with any of the three types already mentioned it seems to be beyond the scope of Panther and Thornburg’s typology.

In another proposal, Kövecses and Radden (1998, 1999) have pointed to the existence of a set of ‘metonymy-producing relationships?or generic principles. The recognition of a series of high-level configurations lays bare one of the most basic problems typologies pose: the need to describe metonymy on different levels of abstraction. On the basis of Kövecses and Radden’s (1998, 1999) discussion, we suggest that it is possible to refine Panther’s and Thornburg’s (1999) typology by making a distinction between low-level metonymies and high-level metonymies. Low-level metonymies are those which make use of non-generic idealised cognitive models in their activation. Non-generic ICMs are conventional representations based on experience which specify elements and their properties and relations. Within this group, we distinguish two further subtypes: propositional and situational. Propositional metonymies correspond to typical cases of metonymy where a concept stands for another in a domain-internal relationship. By way of illustration, consider (2):

(2)

(a) Tired faces all of them, some old, some young.

(b) Blufton smiled with pleasure and kept searching for their faces.

In these sentences we find instantiations of the FACE FOR PERSON mapping where ‘face? which is the source domain, is a subdomain of ‘person? Regarding (2a), ‘faces? is used to stand for ‘people? This is possible on the basis of our experience that a person’s tiredness is particularly evident in his face. Similarly, (2b) is based on the fact that the face, over other body parts, provides the easiest way to recognise a person[5].

In a situational metonymy, a highly striking or otherwise significant element of a specific situation is used to stand for the larger event of which it forms part. An example of this metonymy is found in (1) where the partial scene of the dog leaving with its tail between its legs activates the full picture of the dog being punished and leaving in the form described.

High-level metonymies are mappings which make direct use of generic ICMs, which are in turn abstractions over a number of non-generic ICMs. It is this generic character that allows them to operate at non-lexical levels and to underlie several grammatical phenomena (Ruiz de Mendoza and Pérez, 2001). As is the case with low-level metonymies, this group can be further subdivided into propositional and situational metonymies. The former operates in most cases of what may be called grammatical metonymy, i.e., a metonymy which has consequences in terms of linguistic structure (c.f. section 3.2. below). For example, (3) contains the metonymy instrument for action. As a consequence of the metonymy, a shift of category takes place from noun to verb, which, in turn, brings about the reorganization of the clause:

(3) He hammered the nail into the wall.

The latter group subsumes Panther and Thornburg’s (1998, 1999) illocutionary metonymies. This metonymic type is employed in indirect speech acts, that is to say, whenever a part of a speech act scenario is employed metonymically to stand for the whole of it. Panther and Thornburg (1998) have postulated that speech acts are organised as scenarios consisting of three components which they label before, core/result, and after; each of these components can metonymically stand for the whole scenario (see section 4.1.). Consider (4):

(4) Can you open the window?

This sentence, which is a polar question about the ability of the hearer, is easily understood as a request since it forms part of one of the elements of the before component of the request scenario. Hence, a part of the request scenario (i.e. the ability of the hearer) metonymically stands for the whole scenario (i.e. the request). For further information on speech acts metonymies, see Thornburg and Panther (1997) and Panther and Thornburg (1998, 1999).

Apart from their impact on grammar, high-level metonymies play a crucial role in cognition. Thus, the GENERIC FOR SPECIFIC/SPECIFIC FOR GENERIC mappings have been found to lie at the basis of some higher order cognitive processes. The importance of the relationships which hold between ‘generic?and ‘specific?in the organization and processing of information was first noted by Lakoff and Turner (1989). These authors, however, granted these relationships metaphorical status. More recent accounts (Kövecses and Radden, 1999; Panther and Thornburg, 1999) have convincingly argued that the generic/specific distinction is metonymic in nature, ‘specific?being a subdomain of ‘generic? In addition to this observation, we note that the relationship between these two ICMs is not an identifying one but rather of the ‘stand-for?kind. Kövecses and Radden (1999: 34) have already hinted at the importance of these metonymies for the interpretation of proverbs. By way of example, consider the famous saying Too many cooks spoil the broth, where the general understanding of a particular situation is licensed by the SPECIFIC FOR GENERIC metonymy, which is, in turn, applied to a particular situation through the GENERIC FOR SPECIFIC mapping[6].

If we compare Panther and Thornburg’s typology and our proposal, we observe three main differences:

1.- Although Panther and Thornburg (1999, 2000) have posited the existence of high-level metonymies and have studied the behaviour of some of them across languages, they have not applied the high-level/ low-level distinction in their classification. As a result, they mix both levels of description in their taxonomy. On the one hand, referential metonymies belong to the low-level type; on the other hand, predicative and illocutionary metonymies are cases of the high-level type.

2.- Panther and Thornburg do not take into account situational low-level metonymies so their proposal does not account for metonymies like the one exemplified in (1).

3.- Panther and Thornburg argue for the existence of illocutionary metonymies (based on speech act scenarios) and of high-level metonymies (based on generic ICMs). However, since speech act scenarios are in fact generic situational ICMs, illocutionary metonymies can be just as well described as high-level situational metonymies.

As we have previously mentioned, high-level metonymies have been suggested to have an impact on grammatical structure. Thus, Panther and Thornburg (2000) have analyzed the grammatical consequences of the metonymy ACTION FOR RESULT in stative predicates and of the metonymy EFFECT FOR CAUSE in the What’s that N? construction. In the following sections, we shall apply our taxonomy of metonymies to the study of some grammatical phenomena (section 3) and of some conventionalized illocutionary constructions (section 4) in order to provide a more exhaustive description of their nature and motivation.

3. Propositional high-level metonymy and linguistic structure

3.1. Stative predicates

As Panther and Thornburg (2000) have noted, there are some grammatical constructions which, being typically associated with action predicates, can make use of the stative predicate ‘be? as illustrated by the following sentences:

(5)

(a) Imperative: Be quiet.

(b) I want to know how to: I want to know how to be rich.

(c) Why not: Why not be sincere?

(d) What about: What about being sincere?

The reason why we can employ a stative predicate in typical actional constructions is that underlying these examples there is a source-in-target metonymy which we can call RESULT FOR ACTION. As a consequence of the activation of this metonymy, (5a) is to be interpreted as ‘act (intentionally) in such a way that, as a result, you will be quiet? In our view, the RESULT FOR ACTION metonymy also accounts for why there are some asymmetries in the use of such constructions with stative predicates. For example, (6) is possible whereas (7) is not:

(6) What about being happy?

(7) * What about falling asleep?

The unacceptability of (7) arises from the fact that neither the speaker nor the hearer have the capacity to intentionally get the state of affairs described by the sentence to obtain. However, the imperative negative construction is possible as in Don’t fall asleep. Here the hearer is cautioned to act in such a way that he will not fall asleep. However, falling asleep intentionally is something beyond our control, which makes (7) unacceptable. As is evident, all the ingredients of the matrix domain are necessary for the interpretation of expressions instantiating actional constructions.

3.2. Recategorization of nominal and verbal predicates

The generic character of the action ICM gives rise to a large number of metonymies, many of which underlie grammatical phenomena (probably, the most productive ones are AGENT FOR ACTION, ACTION FOR INSTRUMENT, ACTION FOR AGENT; see Kövecses and Radden, 1998, 1999, for a comprehensive list of metonymies based on the action ICM). Consider the following sentence:

(8) Ian swiftly chested the ball.

In this example, underlying our understanding of the verb “to chest?is the source-in-target metonymy INSTRUMENT FOR ACTION. “To chest?profiles the instrument used by Ian to hit the ball; its meaning may be contrasted with the more generic meaning of the verb “to hit?in which information about the instrument is not provided (cf. Ian swiftly hit the ball with his chest vs. ??Ian swiftly chested the ball with his chest). The full significance of “to chest?is a direct consequence of the source-in-target nature of the underlying mapping whose target is the matrix domain in its entirety.

We may compare this situation with the one found in relation to the semantic analysis of the deverbal noun “cut?in a deep cut, underlying which is the target-in-source metonymy ACTION FOR RESULT. This metonymy, by highlighting a subdomain of the propositional action ICM, involves the reduction of the conceptual material which is brought to bear upon interpretation. In both cases, we have a recategorization of the word class, which has syntactic consequences: verbs and nouns behave differently within the clause. We can compare the case of deverbal nouns and denominal verbs obtained through metonymy with what Halliday (1994) has termed grammatical metaphor. A grammatical metaphor is the result of the grammar of language allowing parts of the system to be expressed in a non-congruent form. Processes are congruently expressed as verbs, but they may be reworded metaphorically as nouns, which designate participants in a process. This is the origin of nominalizations like “argument? which is derived from “argue? Our theory of mappings allows us to understand the difference between nominalizations based on grammatical metaphor and deverbal nouns obtained through metonymy. A metaphor is a domain-domain mapping, which allows us to preserve the original actional structure of the verb in the nominalization. A deverbal noun, on the other hand, is based on a domain-subdomain relationship, so we only make use of a relevant part of the original actional structure of the verb. However, since both processes have much in common, the metonymic mechanism which results in the recategorization of a word may be aptly called grammatical metonymy.

3.3. Valency extension and reduction

It is a well-known fact that it is possible to make intransitive uses of typically transitive verbs, as well as transitive uses of intransitive predicates as is illustrated in (9):

(9)

(a) The door closed.

(b) John walked the dog.

The deletion of an argument of a predicate is called by Dik (1989, 1997) valency reduction whereas the converse operation, in which one extra argument is added to a typically intransitive structure, is called valency extension. Since the two phenomena (i) exploit a generic ICM (the action frame), (ii) have obvious syntactic consequences, and (iii) involve a change in the predicate type, they are likely candidates for metonymic motivation. This is indeed the case. As Ruiz de Mendoza and Pérez (2001) have pointed out, (9a) is an example of the metonymy ACTION FOR PROCESS, and (9b) of the metonymy ACTIVITY FOR THE EVENT CAUSED BY IT.

There is another group of changes involving intransitivization which responds to the metonymy ACTION FOR (ASSESSED) RESULT. Consider (10):

(10)

(a) This bread cuts easily.

(b) *This bread cuts

Sentence (10a) does not focus on the action but on the result of the action. Note additionally that it is difficult to use “cut?without assessing the result of the action as evidenced by (10b). The evaluative adverb serves as an indication that the metonymy ACTION FOR (ASSESSED) RESULT, rather than ACTION FOR PROCESS, is to be invoked. However, in the absence of such an indication, the default reading of a transitive verb used intransitively tends to make use of the ACTION FOR PROCESS metonymy as in (9a), unless this is conceptually incoherent, as in (10b). Note that while it is possible to imagine a door closing without thinking of an intentional agent (e.g. because of the wind), the presence of an intentional agent is necessary in the case of the bread being cut.

3.4. Argument structure

An argument can be defined as the structure formed by a noun (or a pronoun), designating an entity, plus a number of modifiers which range from adjectives and defining relative clauses, to articles, demonstratives, quantifiers, and classifiers. Here we discuss metonymic constraints on (i) the subcategorial conversion of nouns; (ii) the recategorization of adjectives; (iii) the role of some non-adjectival head modifiers.

Although Dik (1989) has carried out an exhaustive account of the subcategorial conversion of nouns, he does not attempt to find a motivation for this frequent phenomenon. We suggest that the motivation is metonymic in nature. Take the following examples from Dik (1989: 121):

(11)

(a) There is too much chair in this room.

(b) There were three Johns at the party.

(c) I would like three butters please.

In (11a), we find a Count > Mass conversion, with the underlying metonymy OBJECT FOR MATERIAL CONSTITUTING THAT OBJECT; similarly, in (11b), where a Proper > Count Noun conversion takes place, we have AN (INDIVIDUAL) ENTITY FOR A COLLECTION INCLUDING THAT ENTITY; and in (11c) which involves a Mass > Count conversion, the metonymy is MATERIAL FOR ENTITY CONTAINING/ HOLDING THE MATERIAL.

These metonymies which only perspectivize a nominal type from different angles do not seem to have syntactic consequences. On closer inspection, however, we observe that at least the metonymy OBJECT FOR MATERIAL may be syntactically relevant in that it motivates a grammatical construction, which not only indicates an excess situation but also the speaker’s negative attitude with respect to it, as shown in (12):

(12) There’s too much superficial starlet biography in the market.

Note that a time or place satellite (or adjunct) is needed for sentences (11a) and (12) to be possible (cf. ?There is too much chair, ?There is too much biography). In combination with the constant character of the meaning implications mentioned above, this observation strongly argues for the status of ‘too much + count noun + adjunct?as a grammatical construction in its own right.

The second type of mapping found in the argument structure is labelled A DEFINING PROPERTY FOR AN ENTITY. This metonymy recategorizes adjectives into nouns, (e.g. blacks for ‘black people?. It has a reverse version in AN ENTITY FOR ONE OF ITS PROPERTIES, which is exemplified in There is a lot of America in what she does, where by “America?we mean stereotypical American values and life style. This is a target-in-source metonymy and involves no recategorization but only subcategorial conversion (Proper Noun > Mass noun).

The analysis has shown that recategorization generally occurs when there is a gap between the categories which express the source and target domains, that is to say, the metonymic target designates an entity which is normally expressed by a category different from the one which realizes the source. It is immaterial whether the matrix domain is the source or the target of the mapping. Thus, in AGENT FOR ACTION, while an agent is realized by a noun, the action, which is both the matrix domain and the metonymic target, is expressed by a verb. In ACTION FOR RESULT, the action, which is the matrix domain and the metonymic source, is realized by a verb but the metonymic target takes the form of a noun.

However, the picture changes dramatically when we are dealing with shifts involving properties. In recategorization rather than in subcategorial conversion, the target of the metonymy needs to be the matrix domain, too. In the source-in-target metonymy A DEFINING PROPERTY FOR AN ENTITY, there is a recategorization of the source since the target domain, which is the matrix domain, designates an entity. In the target-in-source metonymy AN ENTITY FOR A PROPERTY, in contrast, the target domain (which would normally be expressed by an adjective) is not the matrix domain, which rules out recategorization and only calls for semantic reinterpretation of the nature of the source.

Finally, we explore the role of non-adjectival head modifiers in cases of argument structure involving metonymy. Compare (13) and (14):

(13) I prefer Goya to Veláquez

(14) I can’t understand Schliegel

At first sight, we may easily think that we have two simple realizations of the AUTHOR FOR WORKS metonymy. A closer look, however, reveals that there are linguistic constraints on the activation of this metonymy, which are evident from the differences that exist between the following pairs of examples:

(15)

(a) She has a Degas in her studio.

(b) ?I have a Marlowe on the table.

(16)

(a) ?Degas hangs in the hallway.

(b) Marlowe is on the table.

In contrast to (13) and (14), which have a generic reading, the metonymies in (15) and (16), which involve the presentation format of the type of work, have a specific reading. They are based on a double metonymic mapping consisting of the combination of a target-in-source metonymy plus a source-in-target-one; we label this metonymic complex AUTHOR FOR WORKS FOR SAMPLE; the author and the sample are both matrix domains which share the work as a common subdomain[7]. Furthermore, if the sample is unique (the Degas examples), the article is required as a modifier. These examples allow us to see that the scope of argument modifiers carrying a specificity or genericity element is always the least immediate of the target domains involved in the combined mapping. The presence of a genericity element in argument structure, even though it may not be realized morphologically, only triggers off one metonymic mapping, while the presence of a specificity element requires a double metonymic mapping.

3.5. The predication

Dik (1989) has defined the predication as the result of inserting a number of arguments (or terms) in a predicate frame which specifies the restrictions of such an operation. Ruiz de Mendoza and Pérez (2001) have noted that although some verbs like “enjoy?and “begin?select for an activity (c.f. (17)), which is a very generic concept, sometimes they may also take a non-actional complement as (18) shows:

(17) She enjoyed/began the dance.

(18) He enjoyed/began the soup.

This phenomenon has been studied by Jackendoff (1997: 61) who has described these examples as cases of enriched composition, where the hearer needs to look into the world knowledge structure of the complement for an extension of it which is compatible with the generic complementation requirement of the verb: ‘He enjoyed/began to cook/heat/eat, etc. the soup? Ruiz de Mendoza and Pérez (2001) have argued that enriched composition is to be understood as a phenomenon of contextual parametrization of the unrealized generic value of a generic selection restriction. Underlying this process there is a metonymic source-in-target mapping where the source (i.e. soup) is an entity and the target a specific action (determined by the context) where this entity has a role. We may describe this as AN OBJECT FOR AN ACTION (IN WHICH THE OBJECT IS INVOLVED). This metonymy applies to verbs which select for an activity in general, such as “choose? “finish? “miss? “try? and “want?(e.g. He chose/wanted to cook, eat, etc., the soup). A common feature of these verbs is that they do not express an activity themselves but rather the speaker’s attitude or way of acting with respect to a certain activity; in consequence, the construction may be formulated as ‘Speaker’s Attitude/Mode of Action V + (typically) non-actional NP?

An apparently similar construction is found in what we call ‘do + (typically) non-actional NP? which is illustrated in (19):

(19) This week, he’ll do the carpet and I’ll do the dishes.

This construction suggests that the agent carries out an activity which would normally be expected of him in a given context. As a result, if the complement of ‘do?does not specify the nature of the activity (e.g. ‘hoovering?, it is necessary to find a value for ‘do?which gives us the kind of specification we need. Usually, we find default values (e.g. ‘do the dishes?= ‘wash up the dishes? but specific contexts may provide us with other interpretations. For example, in a situation in which a couple is packing because they are moving to a new house, I’ll do the dishes means ‘I’ll pack the dishes? In this construction what is parametrized is not the activity which the verb selects for but the verb itself. The underlying metonymy is GENERIC FOR SPECIFIC.

The target-in-source GENERIC FOR SPECIFIC metonymy is also operative in the parametrization of generic ‘what is?questions:

(20)

(a) What is John?

(b) What’s that building?

(20a) is a way of asking about John’s job (e.g. ‘a teacher? or his role (e.g. ‘the leader?; (20b) is a question about the identity of the building (e.g. ‘the Royal Palace? or its kind (e.g. ‘a palace?. Besides, Panther and Thornburg (2000) have noted that ‘what is?constructions may respond to the EFFECT FOR CAUSE metonymy. This metonymy is of the source-in-target kind. Take (21):

(21) What’s that noise?

(21) may be used to ask about the origin of the noise (an appropriate answer would be “a burglar trying to break into your house?. Thus, an answer which describes the noise (e.g. It is a high pitched noise) will be found irrelevant or not appropriate. Finally, we have another group of ‘what is?questions whose metonymic grounding is the double metonymic mapping WHOLE FOR PART FOR WHOLE where the initial source and the final target are both matrix domains:

(22) What’s that picture?

An appropriate answer to (22), if asked while pointing to a photograph of Big Ben, could be It’s Big Ben or It’s London, but less likely ?That picture is Big Ben/London. The oddity of repeating the noun phrase anaphorically is explained by the fact that there is a metonymic shift in the interrogative utterance from ‘picture?to the ‘image in the picture? The image can be straightforwardly identified with Big Ben or still be part of another (source-in-target) metonymic shift to London.

3.6. Modality

Dik (1989) has distinguished three types of modality: (i) inherent modality, which accounts for a participant’s ‘ability?and ‘willingness?towards the state of affairs in which he is involved; (ii) objective modality, which deals with the speaker’s evaluation of the likelihood of occurrence of a state of affairs (in terms of certainty or obligation); (iii) subjective modality, which expresses the speaker’s personal commitment to the truth of what he says. Halliday (1994: 357) has further refined the concept of objective modality by posing a distinction between epistemic modality (or modalization), which signals probability (‘may be? and usuality (‘sometimes?, and deontic modality, which conveys either obligation (‘is wanted to? or inclination (‘wants to?. This latter subdivision of deontic modality into obligation and inclination is of special interest to us since it corresponds with certain metonymic phenomena which underlie the value of some deontic expressions. Consider the following sentences:

(23)

(a) You must tidy your room.

(b) I must speak to you, please.

These two examples instantiate two typical uses of the modal auxiliary “must?to express an obligation. But, whereas (23a) involves a real obligation imposed by the speaker to the hearer, in (23b) the obligation comes from the speaker to himself. This latter kind of obligation is normally understood as a desire to carry out the action involved in it, which can be paraphrased as ‘I want to speak to you, please? This interpretation is motivated by the metonymy OBLIGATION FOR DESIRE, which allows us to achieve the full meaning of these expressions. Besides, the existence of this metonymy tends to support Halliday’s refined version of deontic modality.

Another metonymic mapping that we find within inherent modality is Panther and Thornburg’s (1999) metonymy POTENTIALITY FOR ACTUALITY which is intantiated in (24):

(24) I can see the mountains from the balcony of my room

In this example the speaker refers to the fact that he actually sees the mountains and not to his ability to see them. The logic which lies at the basis of this source-in-target metonymy is that in order to actually perceive something, one must have the ability to do so. The POTENTIALITY FOR ACTUALITY mapping is very productive in the domain of physical and mental perception, but it can also be found when the actor commits himself to some course of action or personally guarantees the truthfulness of what he says. We suggest that this connection is motivated by the fact that perception is usually taken as evidence (and therefore a guarantee) of factuality:

(25) We can assure (= we assure) you that’s our best guide.

4. Situational high-level metonymies and linguistic structure

4.1. Indirect speech acts

As explained in section 2 above, speech acts can be described in terms of scenarios which contain three main parts: (a) the before, which encompasses all the preconditions; (b) the core, which includes the essential conditions, plus the result, which contains the immediate consequences; (c) the after, which refers to the non-necessary consequences. Consider (26):

(26)

(a) Will you marry me?

(b) It’s late. I can give you a lift.

The first of these two sentences takes the form of a question and the second of a statement; however, (26a) is normally interpreted as a request and (26b) as a proposal. These readings are motivated in each of the two examples by a metonymic mapping which allows us to activate a whole speech act scenario by having access to just one of its parts. Example (26a), which focuses on a prospective future action by the hearer, exploits the after component of the request speech act, while example (26b) focuses on the ability of the speaker (i.e. an element of the before component of the commissive speech act scenario). The underlying source-in-target metonymy explains why it is possible to use a peripheral part of the scenario to refer to the whole of it and accounts for the easiness with which the hearer interprets them (see figure 1 and 2). Furthermore, the grammaticalization of the two constructions exemplified in (26a) and (26b) is based on metonymy. Finally, it should be borne in mind that illocutionary scenarios are abstractions over specific scenarios. Thus, throughout our lives we have seen countless situations where a request takes place; it is by abstracting away all the common elements of these situations that the request scenario, a generic situational ICM, is created.

REQUEST TO PERFORM

Target AN ACTION (Request scenario)

FUTURE Source

ACTION (After component)

Figure 2: A FUTURE ACTION FOR THE REQUEST TO PERFORM THE ACTION.

COMMITMENT TO PERFORM

Target AN ACTION (Commissive scenario)

ABILITY TO PERFORM AN Source

ACTION (Before component)

Figure 3: ABILITY TO PERFORM AN ACTION FOR COMMITMENT TO PERFORM THE ACTION.

4.2. The What’s X doing Y? construction

The intricacies of the What’s X doing Y? construction have been analysed by Kay and Fillmore (1999), who have tried to unravel its peculiar nature. In our view, the semantic value of this construction can be easily explained by positing the existence of an underlying metonymy. Consider the following humorous dialogue taken from Kay and Fillmore (1999: 4):

(21)

A: Waiter, what’s this fly doing in my soup?

B: Madam, I believe that’s the backstroke.

The punchline in (21) lies in the inadequacy of the waiter’s answer. This occurs because the customer’s question is not to be understood as a request for information, but as a complaint. We postulate, in accordance with our theory of generic scenarios, that this meaning is derived on the basis of the metonymy ACTION FOR (ASSESSED) RESULT, which is of the target-in-source kind, and has the generic action ICM as its source domain. Note that doing is a compulsory element of this construction whose function is to activate the action ICM. This ICM takes the form of a generic scenario which can be partially described as follows: (a) actions are controlled attempts to modify a state of affairs, (b) actions have results which can be harmful for the participants and (c) whenever the results of an action are not beneficial for one or more participants, other participants should do their best to change this situation (cf. Ruiz de Mendoza, 1999b for details).

In the realization of the ACTION FOR (ASSESSED) RESULT metonymy in (21) the potential negative consequences of an action are highlighted by questioning the action. In this case, the nature of the metonymic target is cued by the logic of the action ICM which shows that the consequences of an action are a significantly prominent subdomain of the model. In contrast, in non-situational low-level metonymies of the target-in-source kind, it is usually the predicate of the expression (sometimes reinforced by contextual clues) that guides the hearer in the activation of the relevant target (e.g. in Bush attacked Iraq, the predicate ‘attacked?cues the metonymic shift from ‘Bush?to ‘the US army?.

In many contexts, the What’s X doing Y? construction is interpreted as a request to change the state of affairs which troubles the speaker. This is done in accordance with part (c) of the action ICM, and it requires a second metonymic mapping as illustrated in figure 4 below. Consider (22) in this connection:

(22) What’s John doing in the garden?

First, an ACTION FOR (ASSESSED) RESULT metonymy takes place which makes us interpret (22) as ‘John is doing something wrong in the garden? Second, there is another mapping by means of which this sentence is viewed as a request of the type Stop someone from doing something. In this second metonymy, ‘John is doing something wrong in the garden?is part of the before component of the request scenario and metonymically stands for the whole scenario.

ACTION

RESULT/

BEFORE COMP.

REQUEST SCENARIO

Figure 4: The metonymic motivation of the WHAT’S X DOING Y? construction

5. Conclusion

In this paper we have proposed a typology of metonymic mappings based on two criteria. One pays attention to the need to describe metonymy at different levels of abstraction and allows us to distinguish two main metonymic types: high-level metonymies, which make use of generic ICMs, and low-level metonymies, which are based on non-generic ICMs. A second criterion focuses on the ontological nature of the domains involved. Following this criterion, we make a distinction between propositional and situational metonymies. Then we observe that only high-level metonymies interact with grammar. On the one hand, propositional high-level metonymies underlie several grammatical phenomena such as categorial and subcategorial conversions, and help to create specialized constructions whose value can only be fully apprehended with reference to their underlying metonymic mapping. On the other hand, situational high-level metonymies have been found to account for the shifts of illocutionary force conventionally associated with some constructions. In general, our analysis tends to support, at the grammatical level, the analysis carried out at the conceptual level in previous studies.

6. References

Barcelona, A. 2000. “Types of arguments for metonymic motivation of conceptual metaphor? Barcelona, A (ed.). Metaphor and Metonymy at the Crossroads, Berlin/ New York: Mouton de Gruyter. 31-58.

Croft, W. 1993: “The role of domains in the interpretation of metaphors and metonymies? Cognitive Linguistics 4-4: 335-370.

Díez Velasco, Olga I. 2000. “A cross-linguistic analysis of the nature of some hand metonymies in English and Spanish? Atlantis 22-2: 51-67.

Díez Velasco, Olga I. 2001. “Body part metonymies in action and perception frames: a cognitive analysis? EPOS; forthcoming.

Dik, Simon C. 1989. The Theory of Functional Grammar, Vol.1: The Structure of the Clause. Dordrecht: Foris.

Dik, Simon C. 1997. The Theory of Functional Grammar, Vol. 2: Complex and Derived Constructions. Edited by K. Hengeveld. Berlin/ New York: Mouton de Gruyter.

Dirven, Rene. 1993. “Metonymy and metaphor: different mental strategies of conceptualisation? Leuvense Bijdragen 82: 1-25.

Fauconnier, Gilles and Turner, Mark. 1995. “Conceptual integration and formal expression? Metaphor and Symbolic Activity 10: 183-204.

Gibbs, Raymond. 1994. The Poetics of Mind. Figurative Thought, Language, and Understanding. Cambridge: Cambridge University Press.

Goossens, Louise. 1995. “Metaphtonymy: the interaction of metaphor and metonymy in expression for linguistic action? Goossens, L. et al. (eds.) By Word of Mouth. Metaphor, Metonymy and Linguistic Action in a Cognitive Perspective. Amsterdam/ Philadelphia: John Benjamins. 159-174.

Halliday, Michael A. K. 1994. An Introduction to Functional Grammar. London: Edward Arnold. 2nd. ed.

Jackendoff, Ray. 1997. The Architecture of the Language Faculty. Cambridge, Mass.: The MIT Press.

Kay, Paul and Fillmore, Charles J. 1999. “Grammatical constructions and linguistic generalizations: The What’s X doing Y? construction? Language 75, 1: 1-33.

Kövecses, Zoltán and Radden, Günter. 1998. “Metonymy: developing a cognitive linguistic view”. Cognitive Linguistics 9, 1: 37-77.

Kövecses, Zoltán and Radden, Günter. 1999. “Towards a theory of metonymy? Panther, K. and Radden, G. (eds.) Metonymy in Language and Thought. Amsterdam/ Philadelphia: John Benjamins. 17-60.

Lakoff, George and Johnson, Mark. 1980. Metaphors We Live By. Chicago/ London: University of Chicago Press.

Lakoff, George and Turner, Mark. 1989. More Than Cool Reasons: A Guide to Poetic Metaphor. Chicago: The University of Chicago Press.

Lakoff, George. 1987. Women, Fire, and Dangerous Things: What Categories Reveal About the Mind. Chicago: University of Chicago Press.

Langacker, R. W. 1993. “Reference-point constructions? Cognitive Linguistics 4-1: 1-38.

Panther, Klaus-Uwe and Thornburg, Linda. 1998. “A cognitive approach to inferencing in conversation? Journal of Pragmatics 30: 755-769.

Panther, Klaus-Uwe and Thornburg, Linda. 1999. “The Potentiality for Actuality metonymy in English and Hungarian? Panther, K. and Radden, G. (eds.). Metonymy in Language and Thought. Amsterdam/ Philadelphia: Johns Benjamins. 333-357.

Panther, Klaus-Uwe and Thornburg, Linda. 2000. “The effect for cause metonymy in English grammar? Barcelona, A (ed.). Metaphor and Metonymy at the Crossroads, Berlin/ New York: Mouton de Gruyter. 215: 232.

Peña Cervel, M?Sandra. 2001. “Situational metaphors and metonymy? Paper read at the 11th Susanne Hübner Seminar, University of Zaragoza.

Pérez Hernández, Lorena and Ruiz de Mendoza Ibáñez, Francisco J. 2001. “Conceptual Interaction in the Understanding of indirect directive speech acts?Journal of Pragmatics; in press.

Radden, G. 2000. “How metonymic are metaphors?? Barcelona, A (ed.). Metaphor and Metonymy at the Crossroads, Berlin/ New York: Mouton de Gruyter. 93-108.

Ruiz de Mendoza Ibáñez, Francisco. J. 1997a. “Metaphor, metonymy and conceptual interaction? Atlantis Vol.19, 1: 281-295.

Ruiz de Mendoza Ibáñez, Francisco. J. 1997b. “Cognitive and pragmatic aspects of metonymy? Cuadernos de Filología Inglesa, 6/2: 161-178.

Ruiz de Mendoza Ibáñez, Francisco. J. 1999a. “From semantic underdetermination via metaphor and metonymy to conceptual interaction?Essen: LAUD N?492.

Ruiz de Mendoza Ibañez, F. J. 1999b. Introducción a la teoría cognitiva de la metonimia. Granada: Método Ediciones.

Ruiz de Mendoza Ibáñez, Francisco J. 2000. “The role of mappings and domains in understanding metonymy? Barcelona, A. (ed.). Metaphor and Metonymy at the Crossroads. A Cognitive Perspective. Berlin, New York: Mouton de Gruyter. 109-132.

Ruiz de Mendoza Ibáñez, Francisco J. and Pérez Hernández, Lorena. 2001. “Metonymy and the grammar: motivation, constraints and interaction? Language and Communication, Oxford: Pergamon Press; in press.

Thornburg, Linda and Panther, Klaus-Uwe. 1997. “Speech act metonymies? Liebert, W.A. et al. (eds.) Discourse and Perspectives in Cognitive Linguistics. Amsterdam/ Philadelphia: Benjamins. 205-219.

Turner, Mark and Fauconnier, Gilles. 2000. “Metaphor, metonymy and binding? in Barcelona, A. (ed.). Metaphor and Metonymy at the Crossroads. A Cognitive Perspective. Berlin/ New York: Mouton de Gruyter. 77-90.

[1].- Correspondence to Francisco Jos?Ruiz de Mendoza Ibañez and Olga Isabel Díez Velasco, University of La Rioja, Departament of Modern Languages, Edificio de Filología, c/ San Jos?de Calasanz s/n, 26004, Logroño, La Rioja, Spain; tel. 941299430; fax 941299419; e-mail: franruizdfm.unirioja.es. Financial support for this research has been provided by the DGES, grant no. BFF2000-0934.

[2].- Preliminary work in this connection has been carried out by Panther and Thornburg (1999, 2000), Ruiz de Mendoza and Pérez (2001), and Díez (2001).

[3].- Because of the intentionally programmatic character of the present study, we shall illustrate our discussion with examples from English. The need for cross-linguistic analysis of high-level metonymy is well evidenced in Panther and Thornburg’s (1999) study of the POTENTIALITY FOR ACTUALITY metonymy in English and Hungarian.

[4].- If sentence (1) were uttered making reference to a person instead of an animal, we would be dealing with a case of conceptual interaction between metaphor and metonymy. For further information on interactional patterns between metaphor and metonymy, see Goossens (1995); Ruiz de Mendoza (1999a).

[5].- Lakoff (1987) notes that this metonymy works actively in our culture as the tradition of portraits shows. Panther and Thornburg (1998) also observe that passport photos, which have a clear identifying function, only include this body part.

[6].- Peña (2001) has contended that these metonymies have a key role in the creation of what Fauconnier and Turner (1995) have called generic spaces in their many-space model of conceptual interaction. Fauconnier and Turner (1995) argue that the interpretation of metaphor needs, at least, the activation of four different mental spaces. A minimum of two input spaces is projected to another space (the blended space). A fourth space (the generic space), which contains basic skeletal structure derived from the source and target inputs, licenses the projection. Peña (2001) argues that a SPECIFIC FOR GENERIC metonymy underlies the abstraction of information from the source input to the generic space, while the converse metonymy GENERIC FOR SPECIFIC is needed to project the information from the generic space to the target input.

[7].- An easy way to observe the existence of two metonymic mappings in (15) and (16) is provided by the test of anaphoric reference (cf. section 1). Thus, in Marlowe is on the table, it?I>s bound in leather, the anaphoric pronoun makes reference to a sample of Marlowe’s work in book format (i.e. the second matrix domain); on the other hand, in Marlowe is on the table; he’s not easy to read, it is the initial matrix domain that provides the antecedent for the anaphoric pronoun. Choosing between the initial and the final matrix domains for anaphoric reference is simply a matter of conceptual coherence with the rest of the predication.

Rachel Giora and Argyris K. Stringaris

The interest in how the brain processes METAPHORS traces its origins back to a tradition which regarded figurative language as POETIC and hence the opposite of literal language. Despite its ubiquity (Lakoff and Johnson 1980), the underlying assumption has been that this difference should be reflected both in behavioral (Grice 1975; Searle 1979) and brain mechanisms. In this chapter we examine this and other long-standing assumptions, suggesting that the interactions of linguistics with empirical, neuropsychological, and neuroscientific research have drawn a far more complex and, arguably, fascinating picture, not only about metaphor but also about the brain.

Is Metaphor Really So Different?

Since the 1970s, the assumption that metaphors are processed differently from literals has come into close scrutiny. For example, on the basis of psycholinguistic experiments, it has been argued that, in the presence of rich and supportive context, metaphors and literals are processed along the same routes (Gibbs 1994; Ortony et al. 1978).

Although some metaphoric and literal expressions require similar processes (Glucksberg 2001), it has also become increasingly evident that the categories used are in themselves heterogeneous. For instance, some literals (the ring was made of tin, with a pebble instead of a gem) require more complex (metaphor-like) conceptual mapping processes than others (That stone we saw in the natural history museum is a gem; Coulson and Van Petten 2002). Others (curl up and dye) are more appealing although harder to process than metaphoric equivalents (curl up and die; Giora 2003). Metaphors are not all alike either: some are novel, having nonsalient metaphoric interpretations that are usually more appealing yet harder to process than those that are conventional and salient (Giora et al. 2004). Furthermore, some metaphoric stimuli, although relatively conventional, may still be more open-ended than others and, when functioning as a context, give rise to a wider range of associations (Stringaris et al. 2006).

In fact, recent findings indicate that notions such as degree of salience, complexity, or open-endedness may be more suitable to describe the complexity of some of the phenomena in question and span the metaphor-literal divide. Furthermore, whilst these notions may, to an extent, overlap, none of them is specific to metaphor.

Is Metaphor Processed Differently in the Brain?

Consistent with the prevailing view of the RIGHT HEMISPHERE (RH) as being more adept at CREATIVITY than the LEFT HEMISPHERE (LH), early lesion studies have been interpreted as evidence that metaphors rely more heavily than their literal counterparts on regions in the RH (Winner and Gardner 1977). However, Winner and Gardner’s study actually reveals that patients with RH lesions were “not insensitive to metaphor” (p. 725) when offering verbal explications to figurative stimuli, although they tended to erroneously select literal over metaphoric interpretations in a picture matching task. Similarly, the results of the earliest imaging study in the field (Bottini et al. 1994) were also seen as supporting a RH predominance for metaphor comprehension. However, alternative explanations may be more appropriate, given that the linguistic items used also differed on categories other than sensu strictu metaphoricity.

Indeed, subsequent studies have challenged the purported predominance of the RH by demonstrating that, when processing conventional metaphors compared to literals, the LH is more active (Ahrens et al. 2007; Lee and Dapretto 2006; Oliveri et al. 2004), perhaps reflecting retrieval from SEMANTIC stores. In fact, most recent research suggests that, in the absence of a rich biasing context, the hemispheres are insensitive to figurativeness. Rather, the RH is more sensitive than the LH to novel, non-salient interpretations and poetic associations, to complexity, and to open-endedness (Blasko and Kazmerski 2006; Giora 2007). This is corroborated by a recent FMRI study, showing that failure to recruit RH areas when processing novel metaphors distinguishes patients with schizophrenia from healthy controls (Kircher et al. 2007).

Taken together, these findings suggest that LATERALIZATION in the brain’s hemispheres is contingent upon factors such as novelty, semantic and conceptual MAPPING complexity, and evoked range of associations, all of which seem to act independently of figurativeness, thus challenging as too simplistic the notion of a preferential RH processing of stimuli solely by virtue of their metaphoricity. These factors, however, are in accordance with an alternative account – the fine-coarse semantic coding hypothesis (Beeman 1998; Jung-Beeman 2005) – which views the LH as adept at processing finely-tuned semantic relations and the RH as specialized in processing distant semantic relationships.

Novelty

Recent studies indicate that the degree of novelty of an expression is an important determinant of neural processing. For instance, lesion studies (Giora et al. 2000; Kaplan et al. 1990), studies of individuals with Alzheimer’s disease (Amanzio et al. in press), as well as fMRI studies involving healthy participants (Eviatar and Just 2006) demonstrated that processing non-salient (IRONIC, METAPHORIC) interpretations relied more heavily on the RH; processing conventional (metaphoric) meanings involved the LH. Similarly, a series of fMRI, DVF, and ERP studies demonstrated increased activation of RH areas during processing of nonsalient interpretations of novel metaphors (Arzouan et al. 2007; Faust and Mashal 2006; Mashal and Faust in press; Mashal et al. 2005; Mashal et al. 2007) and literal/compositional interpretations of IDIOMS (Mashal et al. in press). And while RH advantage was demonstrated in processing nonsalient interpretations of novel metaphors during first exposure, repeated exposure benefited the LH (Mashal and Faust 2007).

Complexity

That RH recruitment increases with complex sentences has been demonstrated by a number of studies (Jung-Beeman 2005). This has also been seen as typifying conceptual mapping complexity (Coulson and Van Petten 2002), thus introducing another parameter that may determine processing and operate regardless of metaphoricity. Further work is awaited to establish this.

Range of semantic associations

Range of semantic associations, also termed degree of open-endedness, can be seen as determined by the extent to which a stimulus evokes a wide network of semantic associations (Black 1993). In a fMRI study, Stringaris et al. (2006) showed that deciding that a given probe was unrelated to a previous neutral context triggered activation of frontal RH areas following open-ended (metaphoric) contexts (Some answers are straight) but not following more restricted (literal) contexts (Some answers are emotional). In the case of the open-ended PRIMES, both negative and positive decisions elicited the same neural responses. Indeed, higher degree of open-endedness may lead to increased RH activation, probably because of the evocation of remotely related associations (Jung-Beeman 2005). As shown by Mashal et al. (in press), RH areas were uniquely involved when novel literal interpretations of familiar idioms (involving their familiar idiomatic meanings as well) were deliberated on.

Contextual information

CONTEXTUAL factors involved in processing (such as biasing information, task, mood, or experience) further argue against a specific and invariant brain locus for metaphor (Kutas 2006). They show that recruitment of neural networks depends upon factors other than metaphoricity per se. For instance, in Coulson and Van Petten (2007), RH advantage in processing novel metaphors disappears in the presence of biasing information. In Kacinik and Chiarello (2007), both hemispheres were activated by metaphors, but only the LH-response was context-sensitive, thereby restricting the range of possible alternatives. Conversely, the response in the RH indicated retention of alternatives available for processing. Findings in Rapp et al. (2007) indicate that the type of task is an additional determinant of processing. When participants had to judge the EMOTIONAL valence of connotations, metaphors elicited LH regions, despite their novelty. In Stringaris et al. (2006), familiar metaphors activated RH areas when a coherence judgment was required; however, when a meaningfulness judgment was required, same stimuli evoked LH areas (Stringaris et al. 2007). In Blasko and Kazmerski (2006), it was individual differences in experience that mattered: poets and nonpoets differed in which brain areas were recruited when reading poetry.

In sum, recent research, involving a wide range of methodologies, does not provide support for the long assumed special status of metaphor in language. Instead, it shows that the processing of metaphors in the brain depends on a great number of factors beyond figurativeness.

Works Cited and Suggestions for Further Reading

Ahrens, Kathleen, Ho-Ling Liu, Chia-Ying Lee, Shu-Ping Gong, Shin-Yi Fang and Yuan-Yu Hsu. 2007. “Functional MRI of Conventional and Anomalous Metaphors in Mandarin Chinese.” Brain and Language 100: 163-171.

Amanzio, Martina, Giuliano Geminiani, Daniela Leotta, and Stefano Cappa. In press. “Metaphor comprehension in Alzheimer’s disease: novelty matters.” Brain and Language.

Arzouan, Yossi, Abraham Goldstein, and Miriam Faust. 2007. “‘Brain Waves are Stethoscopes’: ERP Correlates of Novel Metaphor Comprehension.” Brain Research 1160: 69-81.

Beeman, Mark. 1998. “Coarse Semantic Coding and Discourse comprehension.” In Right Hemisphere Language Comprehension: Perspectives from Cognitive Neuroscience, 255-284, eds. Mark Beeman and Christine Chiarello. Mahwah, NJ: Erlbaum.

Black, Max. 1993. “More about Metaphor.” In Metaphor and Thought (2nd ed.), ed. Andrew Ortony. Cambridge, MA: Cambridge University Press.

Blasko, G. Dawn and Victoria A. Kazmerski. 2006. “ERP Correlates of Individual Differences in the Comprehension of Nonliteral Language.” Metaphor and Symbol 21(4): 267–284

Bottini, Gabriella., Corcoran Rhiannon, Roberto Sterzi, Eraldo Paulesu, Schenone, P., Scarpa, P. et al. (1994). “The Role of the Right Hemisphere in the Interpretation of Figurative Aspects of Language: A Positron Emission Tomography Activation Study.” Brain 117: 1241-1253.

Coulson, Seana and Cyma Van Petten. 2002. “Conceptual Integration and Metaphor Comprehension: An ERP Study.” Memory & Cognition 30: 958-968.

Coulson, Seana and Cyma Van Petten. 2007. “A Special Role for the Right Hemisphere in Metaphor Comprehension? ERP Evidence from Hemifield Presentation.” Brain Research 1146: 128-145.

Eviatar, Zohar and Marcel Just. 2006. “Brain Correlates of Discourse Processing: An fMRI Investigation of Irony and Metaphor Comprehension.” Neuropsychologia 44: 2348-2359.

Faust, Miriam and Nira Mashal. 2007. “The Role of The Right Cerebral Hemisphere in Processing Novel Metaphoric Expressions Taken From Poetry: A Divided Visual Field Study.” Neuropsychologia 45: 860-870.

Gibbs, W. Raymond Jr. 1994. The Poetics of Mind. Cambridge: Cambridge University Press.

Glucksberg, Sam. 2001. Understanding figurative language: From metaphors to idioms. New York: Oxford University Press.

Giora, Rachel. 2003. On our Mind: Salience, Context and Figurative Language. New York: Oxford University Press.

Giora, Rachel, ed. 2007. Is Metaphor Unique? Neural Correlates of Nonliteral Language. Brain and Language 100/2.

Giora, Rachel, Ofer Fein, Ann Kronrod, Idit Elnatan, Noa Shuval, and Adi Zur. 2004. “Weapons of Mass Distraction: Optimal Innovation and Pleasure Ratings.” Metaphor and Symbol 19: 115-141.

Giora, Rachel, Eran Zaidel, Nachum Soroker, Gila Batori, and Asa Kasher. 2000. “Differential Effects of Right- and Left-Hemisphere Damage on Understanding Sarcasm and Metaphor.” Metaphor and Symbol 15: 63-83.

Kacinik, A. Natalie and Christine Chiarello. 2007. Understanding metaphors: Is the right hemisphere uniquely involved? Brain and Language 100: 188-207.

Kaplan, Joan A., Hiram H. Brownell, Janet R. Jacobs, and Howard Gardner. 1990. The effects of right hemisphere damage on the pragmatic interpretation of conversational remarks. Brain and Language 38: 315-333.

Kircher, T .J. Tilo, Dirk T. Leube, Michael Erb, Wolfgang Grodd, and Alexander M. Rapp. 2007. “Neural Correlates of Metaphor Processing in Schizophrenia.” NeuroImage 34: 281-289

Kutas, Marta. 2006. “One Lesson Learned: Frame Language Processing – Literal and Figurative – As a Human Brain Function.” Metaphor and Symbol 21: 285-325.

Grice, H. Paul. 1975. “Logic and Conversation.” In Speech Acts. Syntax and Semantics Vol. 3, 41-58, eds. Peter Cole and Jerry Morgan. New York: Academic Press.

Lakoff, George and Mark Johnson. 1980. Metaphors We Live By. Chicago: University of Chicago Press.

Jung-Beeman, Mark. 2005. “Bilateral Brain Processes for Comprehending Natural Language.” Trends in Cognitive Sciences 9: 512-518.

Lee, S. Susan and Mirella Dapretto. 2006. “Metaphorical vs. Literal Word Meanings: fMRI Evidence against a Selective Role of the Right Hemisphere.” NeuroImage, 29, 536-544.

Mashal, Nira and Miriam Faust. In press. “Right Hemisphere Sensitivity to Novel Metaphoric Relations: Application of the Signal Detection Theory.” Brain and Language.

Mashal, Nira and Miriam Faust. 2007. “Repeated Exposure to Novel Metaphors Affects Hemispheric Involvement.” Paper submitted for publications.

Mashal, Nira, Faust, Miriam, Talma Hendler. 2005. The Role of The Right Hemisphere in Processing Nonsalient Metaphorical Meanings: Application of Principal Components Analysis to fMRI Data.” Neuropsychologia 43 (14): 2084-2100.

Mashal, Nira, Faust, Miriam, Talma Hendler, and Mark Jung-Beeman. 2007. “An fMRI Investigation of the Neural Correlates Underlying the Processing of Novel Metaphoric Expressions.” Brain and Language 100: 115-126.

Mashal, Nira, Faust, Miriam, Talma Hendler, and Mark Jung-Beeman. In press. “Processing Salient and Less-Salient Meanings of Idioms: an fMRI Investigation.” Cortex.

Oliveri, Massimiliano., Romero, Leonor, and Costanza Papagno. 2004. “Left but Not Right Temporal Involvement in Opaque Idiom Comprehension: A Repetitive Transcranial Magnetic Stimulation Study.” Journal of Cognitive Neuroscience 16: 848-55.

Ortony, Andrew, Diane L. Schallert, Ralph E. Reynolds and Stephen J. Antos. 1978. “Interpreting Metaphors and Idioms: Some Effects of Context on Comprehension.” Journal of Verbal Learning and Verbal Behavior 17: 465-477.

Rapp, M. Alexander, Dirk T. Leube, Michael Erb, Wolfgang Grodd, Tilo T. J. Kircher. 2007. “Laterality in Metaphor Processing: Lack of Evidence from Functional Magnetic Resonance Imaging for the Right Hemisphere Theory.” Brain and Language 100: 142-149.

Searle, John. 1979. Expression and Meaning. Cambridge, England: Cambridge University Press.

Schmidt, L. Gwen, Casey J. DeBuse, and Carol A. Seger. 2007. “Right Hemisphere Metaphor Processing? Characterizing the Lateralization Semantic Processes.” Brain and Language 100: 127-141.

Stringaris, K. Argyris., Nicholas C. Medford, Rachel Giora, Vincent C. Giampietro, Michael J. Brammer, and Anthony S. David. 2006. “How Metaphors Influence Semantic Relatedness Judgments: The Role of the Right Frontal Cortex.” NeuroImage 33: 784-793.

Stringaris, K. Argyris., Nicholas C. Medford, Vincent C. Giampietro, Michael J. Brammer, and Anthony S. David. 2007. “Deriving Meaning: Distinct Neural Mechanisms for Metaphoric, Literal, and Non-Meaningful Sentences.” Brain and Language 100: 150-162.

Winner, Ellen and Howard Gardner. 1977. “The Comprehension of Metaphor in Brain-Damaged Patients.” Brain 100: 717-729.

Bert Peeters

There can be no doubt that structural linguistics, which flourished half a century ago on both sides of the Atlantic Ocean, lived up to its name: it was structural because it considered languages to be self-contained entities that had either to be shaped into a rigorous structure, or actually possessed a structure which was real and merely waiting to be discovered. There can be no doubt either that transformational grammar, which in its heydays pushed structuralism into quasi-total oblivion, lived up to its name: it was transformational because it posited several successive strata or structures in sentence generation which were linked by means of transformations of all sorts. On the contemporary scene, there can be no doubt that functional linguistics lives up to its name: it attaches a great deal of importance to the way in which languages function and to the functions of language. The question that will be raised in the next few pages is the following: does Cognitive Linguistics, as we know it today, live up to its name?

Before I answer the main question, let me ask, and answer, another one. I mentioned structural linguistics, transformational grammar, and functional linguistics, and spelled all of them with lower case initials. Why did I use upper case initials when referring to Cognitive Linguistics (and why am I doing it again)? It is certainly not common practice. Langacker (1998: 1), for instance, points out that the “movement called cognitive linguistics [lower case, B.P.] belongs to the functionalist tradition”. He then goes on to add, quite crucially, that “although its concern with cognition hardly makes it unique, the label cognitive is not entirely arbitrary” (ibid.). The subordinate clause indicates why, in my view, the use of upper case initials is warranted. There is a lot of cognitive linguistics going on outside the movement described by Langacker. Generativists in particular have more than once expressed their annoyance regarding what they see as the “misappropriation” of the term by Cognitive Linguists. Their research interests, and that of many others, carry an equal entitlement to identification by means of the label cognitive linguistics. It is an entitlement which, in the current climate, they will find increasingly difficult to claim.3

Instead of pondering the possible implications of the terminological skirmishes that are taking place, let us return to our main question. Does Cognitive Linguistics (with upper case initials) live up to its name? At one stage (Peeters 1998), the answer (or rather, my answer) came much closer to a two-letter word than it does today. I now suspect that after all there is some room around the cognitive science table for Cognitive Linguistics. However, Cognitive Linguists must do their homework first.4

An increased commitment to certain aspects of reality (to be defined below) is likely to result in much-needed closer ties with, and increased visibility in, the cognitive science community at large (in which, it would seem, they have not as yet acquired their rightful place, in spite of the fascinating facts of language which they have been able to unearth). Only when such closer ties obtain will it be possible to change the answer to the question from `not yet’ to `yes’.

I do realise that even a more considered answer such as this one (more considered at least than the one I formulated in Peeters 1998) is likely to raise many eyebrows. Those who are twitching should recall that the Cognitive Linguistics movement as we know it today was born out of polemical opposition to Chomskyan linguistics. Cognitive Linguists, therefore, ought to be able to handle a bit of polemical opposition directed at themselves. Although possibly corrosive, my remarks intend to be constructive. I see my role as that of a gadfly, and hope that those people who are being bitten won’t ache too much.

1. Reflexions on psychological and biological reality

1.1 God’s truth from structural linguistics to the present day

I started off by saying that structural linguistics was structural because it considered languages to be self-contained entities that had either to be shaped into a rigorous (phonological, morphological, possibly lexical) structure, or actually possessed a (phonological, morphological, possibly lexical) structure which was real and merely waiting to be discovered. In his celebrated review of Zellig Harris’ Methods in structural linguistics (Harris 1951), Householder (1952:260) referred to those two ideological positions by means of the labels God’s truth and hocus-pocus.5

The details are as follows:

On the metaphysics of linguistics there are two extreme positions, which may be termed (and have been) the `God’s truth’ position and the `hocus-pocus’ position. The theory of the `God’s truth’ linguists [...] is that a language has a structure, and the job of the linguist is (a) to find out what that structure is, and (b) to describe it as clearly, economically, and elegantly as he can, without at any point obscuring the God’s truth structure of the language. The hocus-pocus linguist believes (or professes to believe – words and behavior are not always in harmony) that a language [...] is a mass of incoherent, formless data, and the job of the linguist is somehow to arrange and organize this mass, imposing on it some sort of structure (which must not, of course, be in any striking or obvious conflict with anything in the data).6

The hocus-pocus position was fairly widespread. Householder himself did not object to what he called “a certain amount of hocus-pocus” (Householder 1952: 261; emphasis added); in fact, he went on to say that in his view “all linguists indulge in it frequently, for fun; and it is the greatest fun of linguistics” (ibid.).

Just under half a century later, a bewildering variety of descriptive frameworks are doing the rounds. Even if the terms are no longer on everyone’s lips, the distinction, widely accepted by the American structuralists of the classical era (as pointed out on 23 March 1999 by Esa Itkonen on Funknet), remains valid.7

At the same time, most (if not all) contemporary linguists firmly believe in the reality or objective existence of the (sub)structures they describe. Unfortunately, this appears to be another instance where “words and behavior are not always in harmony” (Householder). The structures that are put forward in the present day and age by an ever increasing number of often incompatible accounts are so hugely different that they cannot all exist as such in the material that is being described. They are at best “interpretations” of an internal organisation which remains more or less elusive. The map, as is often said, is not the territory. In some forms of linguistics, it may be more complex (contra Hutton, this volume).

One important difference between the fifties and the nineties needs to be highlighted. For many linguists, psychological and biological reality (or at least likelihood) are more important than ever before. Behaviourism, in the crude form in which it had been imported into linguistics by Bloomfield, had little or no such reality value. With that in mind, the terms God’s truth and hocus-pocus may be redefined (and have been).8

In his Funknet posting, Itkonen presented a set of updated definitions:

The `hocus-pocus’ view (without any negative connotations) has been and is (and will be) represented by those who just want to present the facts of a given language [...] in a maximally simple (sic) and general way. Most of the time, this way has or is meant to have NO psychological or biological reality. [...] The `God’s truth’ position (without any either positive or negative connotations) is represented by those who do not merely wish to capture the psychological and/or biological reality, but who actually succeed in doing so, at least to some extent.9

Cognitive Linguistics has come a long and arduous way, but it has an even longer and more arduous way to go: to increase its chances of real integration in and recognition by the cognitive science community at large, it must engage with this new form of God’s truth, i.e. with psychological and with biological reality – the mind as well as the brain – in a way it has not done hitherto.10

1.2 Psychological vs. biological reality

Nobody would want to deny that Cognitive Linguists have made inroads into the area of the mind. They have done so by asking questions relating to psychological reality at large, and in particular to the nature of categorisation, to the issue of storage versus computation, etc. Hence, I am not saying that there is nothing cognitive about Cognitive Linguistics. However, for most Cognitive Linguists, cognitive seems to be synonymous with psychological. This is not the way the term is used, for instance, among cognitive psychologists, whose subject area is of course not “psychological psychology”. For them, cognitive means `pertaining or related to knowledge’. Just as it is wrong to reduce cognition to neurocognition (as I did more or less in Peeters 1998), it is wrong to reduce cognition to psychology. Nonetheless, when it comes to matters of the brain (i.e. biological reality), there is not a lot of interest just yet. There are certainly multiple references in the literature to the “mind/brain”, but that is often as close as one gets to the brain. In fact, mind and brain are vastly different entities: the former is psychological, the latter biological. As I have pointed out elsewhere (Peeters 1996): “The mind is what the brain does for a living”.11

I would be remiss not to mention at this point the work of scholars such as George Lakoff, Paul Deane and Terry Regier. Deane is the author of a very impressive volume called Grammar in Mind and Brain (Deane 1992), and of a paper (Deane 1996) which examines the effects of agrammatic aphasia on neurological support for Cognitive Linguistics. Regier (1996) has shown that spatial relations as expressed in language have no objective existence in the world, but depend directly upon the structure of the human brain. He is a close associate of Lakoff, who, for the last ten years or so, together with cognitive scientists Jerry Feldman, Lokendra Shastri, David Bailey and Srini Narayanan, has been working at a “neural theory of language”.12

In a recent interview with John Brockman, following the release of Lakoff & Johnson (1998), Lakoff provided the following comment (using the metaphor of neural circuitry defined by Lakoff & Johnson 1998:104, quoted by Jones, this volume; cf. note 10):

A human brain consists of a very large number of neurons connected up in specific ways with certain computational properties. How is it possible to get the details of human concepts, the forms of human reason, and the range of human languages out of a lot of neurons connected up as they are in our brains? How do you get thought and language out of neurons? That is the question we are trying to answer in our lab through the computational neural modeling of thought and language.13

Unfortunately, at this stage, Lakoff remains much better known for his equally important work in metaphor and prototype theory (Lakoff/Johnson 1980, Lakoff 1987, Lakoff/Turner 1989 etc.).14

If there is one Cognitive Linguist who is widely known for having succeeded in capturing the biological reality not just to some extent, but (in my view) to a very considerable extent, and to present that reality, in its full complexity, to his fellow linguists in a relatively easy-to-follow way, it must be Sydney Lamb. According to him (Funknet, 26 March 1999), interest in “God’s truth” can be reformulated for modern times as an interest in what is in the mind and/or the brain. In his newly published Pathways of the Brain (Lamb 1999), which appears set to become a highly controversial work, he provides a fascinating neurocognitive account of the workings of language (and other cognitive abilities). Instead of isolating linguistics from other scientific endeavours, as ignorance of those other scientific endeavours often compels individual scholars to do, he builds bridges to other disciplines. Whether they will stand the test of time, only time can tell.

2. Cognitive Linguistics versus cognitive linguistics

2.1 From cognitive linguistics to Cognitive Linguistics

Just over two decades ago, it was Lakoff, not Lamb, who was taken to task by Chomsky (1979: 150) for “working on `cognitive grammar’, which integrates language with nonlinguistic systems”. Chomsky (1979) is the English version of a text originally published in French in 1977, two years after the term cognitive grammar had first surfaced in Lakoff’s writings (cf. Lakoff/Thompson 1975). Chomsky, for one, did not “see any theory in prospect there”. This flippant remark raises the interesting question of the (hidden) impact which the man from MIT may have had on the Lakoff-Langacker agreement to use a common label for their work (in replacement for Langacker’s term space grammar, which was still in use in the early eighties). In other words, did Chomsky’s criticism backfire? Did Lakoff read Chomsky (1979), and did he think that cognitive grammar was too beautiful a term not to be made use of by Langacker and himself (against Chomsky)?

Whatever the case may be, it is quite ironical that Lamb (1999) felt unable to freely use the term he too had proposed (cf. Lamb 1971), earlier than Lakoff (or Chomsky), earlier even than Lewis Sego, who in late February 1999 reported on Cogling that – I quote – “almost twenty-seven years ago (precisely 13 April 1972), when I synthesized two separate doctoral programs I had nearly completed and therefrom coined the term cognitive linguistics, I considered the underlying concept a possible source of philosophical and scientific peacemaking”.15

Lamb decided instead to resort to the more explicit term neurocognitive linguistics, for fear of being mistaken for one of the many linguists who, by the end of the eighties, had started to use terms such as cognitive linguistics and cognitive grammar in a rather different and much broader way.

Let us recall some of the evidence. Although a preprint had been in circulation since 1984, the year 1987 saw the (official) publication of the first volume of Langacker’s Foundations of Cognitive Grammar (Langacker 1987-1991). Lakoff (1987) was published almost at the same time, and is of course another milestone, even though its title does not refer explicitly to either cognitive grammar or Cognitive Linguistics.16

This was followed in 1988 by a hefty volume on Topics in Cognitive Linguistics (Rudzka-Ostyn 1988), and in 1989 by a gathering in Duisburg (Germany), which was proclaimed to be the first International Cognitive Linguistics Conference. A selection of the papers read during that meeting was published four years later, in a volume (Geiger/Rudzka-Ostyn 1993) which became the pretext for my much-maligned (but also, in other quarters, much-applauded) review article called “Cognitive Musings” (Peeters 1998).17

In other words, what had been, until the end of the eighties, a collocation like any other one, gained the status of a proper name, an ideological label rather than a purely descriptive one, chosen in order to gain legitimacy, and to outdo other cognitivists. The term cognitive linguistics had been around for almost twenty years, but its consistent use as a name to refer to what is today a broadly defined paradigm with a very respectable following was new. It became the name adopted by one particular group of people, led by Lakoff and Langacker, to refer to the sort of work they were undertaking. It also became – quite naturally – the name used by others to identify that particular group of people.

In his interview with John Brockman (cf. section 1.2), Lakoff summarises his contribution to contemporary linguistics as follows:

I set about, along with Len Talmy, Ron Langacker, and Gilles Fauconnier, to form a new linguistics one compatible with research in cognitive science and neuroscience. It is called Cognitive Linguistics, and it’s a thriving scientific enterprise.18

The question arises whether compatibility with research in cognitive science and neuroscience is enough. I am inclined to think that it is necessary, but not sufficient. The trouble for those who think that the work they are undertaking fits Lakoff’s description is that compatibility implies (at least some) awareness. Unfortunately, large numbers of Cognitive Linguists remain unaware of what is happening in cognitive science and especially neuroscience. They practice what Sydney Lamb, in the Funknet posting referred to above, calls analytical linguistics. The latter, he says, is the familiar kind. “In this mode one is mainly concerned with accurately describing linguistic productions (without concern for the process of production or that of comprehension or the system that makes those processes possible)”. Neurocognitive linguistics, on the other hand, aims at “understand[ing] that system and those processes”.19

Geeraerts (1995: 111-112) provides the following useful summary of the sort of work in which most Cognitive Linguists today are engaged:

Because cognitive linguistics [what I call Cognitive Linguistics; B.P.] sees language as embedded in the overall cognitive capacities of man, topics of special interest for cognitive linguistics include: the structural characteristics of natural language categorization (such as prototypicality, systematic polysemy, cognitive models, mental imagery and metaphor); the functional principles of linguistic organization (such as iconicity and naturalness); the conceptual interface between syntax and semantics (as explored by cognitive grammar and construction grammar); the experiential and pragmatic background of language-in-use; and the relationship between language and thought, including questions about relativism and conceptual universals.

A broader use of the label cognitive linguistics than was originally the case (in Lamb’s earlier work) is entirely legitimate, because there is indeed more to cognition than neurocognition. The crucial thing, however, is that in the broader meaning of the label there should be a clearly visible spot for the original use. Unfortunately, for those who were expecting to find them included, neurocognitive issues are conspicuously absent from Geeraerts’ list. They are not among the “topics of special interest” to Cognitive Linguists. And yet, it cannot be denied that the study of neurocognition and of its implications for language is a legitimate part of the study of language and cognition, not something that can be left to a few individuals in the Cognitive Linguistics community (who, by the way, also involve themselves with non-neurocognitive issues).

2.2 The need for more “neurocognitive depth”

More Cognitive Linguistics research is needed in the neurocognitive arena; we must stop thinking that others will do it for us and will by themselves ensure visibility for Cognitive Linguistics outside the narrow bounds of linguistics as a scientific endeavour. Neurocognitive linguistics, the “non-analytical” counterpart to the various “analytical” issues listed by Geeraerts, is an important branch of the overall cognitive enterprise. Sadly, it is hardly even described or referred to in the increasing number of introductions to Cognitive Linguistics which are currently being released. And here, of course, we end up in a vicious circle. As long as only a handful of Cognitive Linguists are doing neurocognitive work, those with enough knowledge of the field to write introductory texts will not find it worthwhile to report on the work of that minority. The introductory texts are being read by newcomers to the field who then, typically, start doing the sort of work that is reported on, without reaching out further.

It is my personal conviction, which I know many others do not share, that the lack of “neurocognitive depth” in Cognitive Linguistics is one reason why cognitive scientists in general do not take more notice of Cognitive Linguistics than they currently do. This is easily verifiable when one inspects the contents of the most commonly used introductions to and readers in cognitive science: whereas some go as far as to ignore linguistics altogether, in any of its forms, others often limit themselves to work which Cognitive Linguists readily dismiss as ill-conceived (e.g. generative grammar, truth-conditional semantics).20

Admittedly, it could be argued that the poor visibility of Cognitive Linguistics in cognitive science texts has an altogether different reason. In his comments on the now superseded abstract which lies at the origins of this paper, Langacker referred to “the utter dominance of generative grammar for several decades and the length of time it realistically takes for a different set of ideas to become known outside (even inside) the field, especially when they depart from the `mainstream consensus’ that outsiders initially look to”. It is quite clear, though, that outside the USA generative grammar is no longer as dominant as it once was. In fact, Cognitive Linguistics itself, and many similarly oriented functional schools, have seriously undermined what used to be an almost unassailable position. There is no such thing as a “mainstream consensus” anymore, and there has not been one for a long time. Finally, whether length of time is an issue can also be questioned: it did not take Chomsky a long time to get noticed by psychologists, and his “set of ideas” was certainly very different from anything that had been heard before. As is well known, he gained early prominence with a lengthy review (Chomsky 1959) in which Skinner’s Verbal Behavior (Skinner 1957) was shred to pieces.21

Skinner, of course, was just about the most respected psychologist of the day. I am reasonably confident that if anyone within the Cognitive Linguistics community were to similarly attack one of today’s outstanding cognitive scientists, the cognitive science community at large would not fail to take notice. And it would not take very long either. But it is not necessarily the best way forward. Chomsky gained prominence among psychologists, less on the merits of his own work (which many found problematical) than on the ferocity with which he destroyed one of theirs, and not just anyone. I do not think that Cognitive Linguists would want to engage in that sort of activity.

3. Whereto from here?

3.1 Innateness and modularity

In March 1999, on Cogling, a step was made in the right direction. A few weeks before, Dick Hudson had asked for suggestions for introductory readings on Cognitive Linguistics for undergraduates. On 10 March, he produced an annotated bibliography which covered not only Cognitive Linguistics, but also innateness and modularity. These were actually the three sections which, in his own words, he had distinguished “rather arbitrarily”. That judgment was wrong inasfar as it is in fact common practice for introductory textbooks in Cognitive Linguistics to refer to innateness and modularity without exploring either at great length (in contrast, there are entire chapters about prototypes, metaphor, frames, or about the traditional disciplines of linguistics such as lexicology, morphology, syntax, phonetics etc.). Hudson’s judgment was however right inasfar as both innateness and modularity are important themes in cognitive linguistics, and should figure much more prominently in Cognitive Linguistics as well. Questions to be asked include the sort of evidence, if any, that can be found in favour or against innateness, in favour or against modularity. Those who provide that sort of evidence (or theoretical justification) are not normally associated with the Cognitive Linguistics movement.22

It is to be hoped that Hudson’s bibliography will actually be put to good use, so that students who learn about Cognitive Linguistics also learn, and read, about innateness and modularity, in more detail than appears to be currently the case.

For completeness, I shall add a few comments on both. Innateness, according to widespread opinion, is a matter of belief rather than of research. It is a theoretical claim. But this, of course, depends on how much you want to assume is innate, the two extremes being the physiology to acquire a language, given appropriate circumstances, vs. an entire universal grammar. The former position is easier to research than the latter, which has indeed defied scientific scrutiny. Even so, evidence is scant and debatable, and the ethical implications of this sort of research loom large. For a recent appraisal, cf. Elman et al. (1996).

Modularity, on the other hand, is a slippery concept as well (Hilferty 2000). In discussions among linguists, the term module is used, not only with reference to language as a whole (as opposed to other presumed modules such as vision), but also – often at the same time – with reference to semantics and syntax, etc., even with reference to components within the latter, all of which according to some can be shown to be separate modules (submodules, sub-submodules). Uriagereka (1999: 268) reminds us that Fodor (1983), who is often identified as the most important catalyst for the recent modularity debate in linguistics, never intended the concept of “module” to be used in that way; he intended it as a… theoretical construct at the higher level (e.g. language, vision), where modularity, like innateness, therefore does appear to remain very much a matter of belief. At lower levels, though, some sort of (limited) modularity does seem to occur, although finding proof of more promises to be a very difficult enterprise.

3.2 Matters of the brain

I shall give another example of what, I believe, also ought to belong in Cognitive Linguistics. I remember watching a television documentary a few years ago, which had been produced in 1994 for the BBC. Its title was The man who made up his mind,23

and it focussed on work by Gerald Edelman. Edelman is not a linguist, let alone a Cognitive Linguist; he is a neurobiologist. But his work on so-called neural Darwinism (Edelman 1987) has implications for language, which I believe the documentary briefly referred to. In Peeters (1998), I report what happened after that. I asked the readership of Cogling for more information on Edelman, thinking Cognitive Linguists would know everything about him that there is to know. The results were contrary to expectation. Two or three replies came in, one of which stated in unambiguous terms that I was asking the wrong people… This left me puzzled. There is hardly anything more cognitive than the question of how language is processed in the brain. For that reason, exploration of brain processes, with special reference to language, ought to be part of the overall brief of Cognitive Linguists.24

In his comments on the abstract of this paper (cf. section 2.2), Langacker wrote as follows:

Maybe we should all study and cite Edelman, but does that tell us how to characterize the meaning of dative case in Polish or describe an antipassive construction? Should we all go work in wetlabs, or can mental spaces and blending be studied without that experience? [...] There has to be a large quantity of work that is specifically linguistic in nature, work that is specifically psychological or neurological, and work that tries to bring these together in one way or another. All are legitimate and important, requiring their own expertise, and they should all be welcomed for their contribution to what is an immense overall investigatory enterprise.

I could not agree more. In fact, we should not “all study and cite Edelman”. But a few more than are currently taking any notice should read him more attentively, in an attempt to understand the implications of his research for our understanding of how language is processed in the brain, and they should tell the rest of us what they have discovered. However, what is really needed is a book of the kind that is now invading computer stores all over the world, a book titled, for instance, Edelman for dummies. Those who have tried to read Edelman will have noticed that he is not an easy author to follow, not even in his so-called popularising account Bright air, brilliant fire (Edelman 1992), in which he approvingly quotes Lakoff, and criticises Chomsky.25

But what if Edelman has got it wrong? Or what if he is not entirely right? The latter stand is taken by Sydney Lamb, who was singled out earlier for his contribution to neurocognitive linguistics. Lamb arrived independently at the basic idea of what Edelman so appropriately calls “neural Darwinism”. In a private e-mail message dated 25 May 1999, he reports that, as he became aware of Edelman’s work, he started to read selectively (as one is often obliged to do these days) and found himself in sufficient agreement to include a few, overall favourable, references in Pathways (Lamb 1999). Later, more exhaustive, readings produced disappointment: having struggled through Edelman (1987), which was a challenging read even to him, Lamb found that Edelman “comes close”, but does not quite understand after all how the brain actually stores information. One is tempted to conclude that Lamb’s exposition of the theory of neural Darwinism may be more accurate than Edelman’s…26

Skeptics who have some idea of what is happening in the brain but do not really want to know more might at this point say the sort of thing that Aya Katz quite aptly expressed on Funknet, on 29 March 1999:

Brain configurations vary. Persons with severe brain damage in early childhood are often capable of normal language processing and production, even though the connections in their brains are very different from the norm.

What if we found that even in normal, undamaged brains, there is an immense variety of ways in which the same item can be stored and processed by native speakers of the same language? If we concentrated on the biological entity that produces it, we’d lose the generalization involved in the communicative function of language.

Speakers don’t know how their interlocutors’ brains are configured. Communication is based on the abstract system of contrasts set up in the language. We react to electronically programmed simulations of human speech just as we would to those produced by actual people, if it’s close enough. We read manuscripts written thousands of years ago, and the information is communicated, even though the brain that produced it has long ago been consumed by worms.

That’s the magic of language. The concretes don’t matter.

There was an immediate reaction from Tony Wright, who later apologised for having somewhat condescendingly referred to “those [linguists] who want to be neurologists” and for having expressed his disbelief that some people would actually “want to give up linguistics for bean counting (neuron-counting)”. Wright knew of course that neurologists do a lot more than just count neurons (otherwise he would not have apologised). He also knew that “the concretes do matter”, even though at the same time he was fascinated by the way Katz had summarised the position of a majority of linguists (including Wright himself) who feel that the brain, for them, is terra incognita.

Why do the concretes matter? First of all, because by looking at language processing in the brain we can subject the very theses that we constantly pay lip-service to but do not pursue any further to detailed scrutiny. We can begin to notice that the oft-repeated statements that we cherish are as extreme as the equally oft-repeated contrary statements heard in other circles. For instance, Cognitive Linguists like to say that there is no separate language module, that language is but one way humans use among several others to interact with the world, using general cognitive mechanisms. Langacker (1998: 1) puts it this way:

[Cognitive linguistics] contrasts with formalist approaches by viewing language as an integral facet of cognition (not as a separate “module” or “mental faculty”). Insofar as possible, linguistic structure is analyzed in terms of more basic systems and abilities (e.g., perception, attention, categorization) from which it cannot be dissociated.

Generativists, on the other hand, are generally ready to swear the exact opposite. A closer look at the brain, and at how things really work, is increasingly likely to show that neither position can be maintained in its extreme form. The truth is somewhere in the middle (Newmeyer 1999). There is a certain degree of modularity, in that language – and in fact every other cognitive mechanism – involves brain activity that is unique to it. But clearly, there is a lot of interaction as well.

The concretes also matter for another reason. There are things that they could teach us, say, about polysemy, which, according to most Cognitive Linguists, is an essential property of a majority of lexical material (see now Cuyckens/Zawada 1999). Polysemy is the norm, rather than monosemy. While I have on various occasions argued against this view, I am now ready to admit that I was looking at things from a purely systematic point of view, without any reference to real language processing. In the meantime, my awareness of storage and computation of language material in the brain has increased – or so I hope -, and I am ready to answer the question: “Is X – where X is a word of whichever language I am studying – polysemous or monosemous?” by saying: “Yes, of course”, i.e. it is both.27

It depends on the speaker. But in order to be totally sure, we might want to check whether study of brain mechanisms is able to enlighten us further. If, for instance, a word which is potentially polysemous were inserted in different disambiguating contexts, and these contexts were read out to subjects whose brain activity is being measured, could we not tell, from the chemical processes and the neuron firings observed, whether that word is more likely to be monosemous (similar firings independent of context) or polysemous (rather dissimilar firings in each context of use)? I am not aware of any work that is being done in this area right now. If there is, we ought to know about it. The aim is to reach an understanding of polysemy which is, in Lamb’s (1999) terminology, not only operationally and developmentally plausible, but also neurologically.28

The “immense variety of ways in which the same item can be sorted and processed by native speakers of the same language” (Aya Katz) is exactly one of the things we must try to understand: viz. why it is possible to have that variety of ways, without significant risk to normal communication being impeded.

Let us take another example. Many Cognitive Linguists undertake research on idealised cognitive models (i.e. cognitive simulations of reality, also known as ICMs). It would be interesting to explore whether, for instance, selling and buying should be associated with one such model (as appears to be current practice) or with two. The shift in simulation perspective may be too significant to stick with just one ICM. It appears to be more significant than the shift observed when an event is being verbalised by means of a passive rather than an active construction, or when a stealing event is looked at from the point of view of the person victimised (X was robbed of Y) rather than from that of the object taken (Y was stolen from X). Coming back to buying and selling, one could think of an experiment where subjects are asked to conceptualise either event, while having their brain activity subjected to detailed observation. I do not know the outcome, but would suggest that, perhaps, there are two models involved rather than a single complex one (or two frames, in Fillmore’s sense; cf. Fillmore 1977, 1982, 1985).29

Someone who buys always buys from a vendor or a salesperson, but someone who sells something does not necessarily sell anything to anyone. He or she may just be trying to sell, be an unsuccessful tradesperson. ??I am buying a house, but there is nobody selling theirs sounds weird in a way that I am selling my house, but I haven’t found a buyer yet does not.

And yes, there are multiple areas of research where so-called “neuron counting” is not going to be of any help. Clearly, the dative case in Polish and antipassive constructions in Australian aboriginal languages and zillions of other phenomena are unlikely to be better understood if we observe what is happening in the brain when they are being uttered or perceived. I must stress once again that not every single Cognitive Linguist is supposed to “study and cite Edelman” (or Lamb for that matter). Similarly, not every single cognitive scientist is supposed to engage in neuroscience. There is a lot of other work to be done, and lots of people are needed to do it. Still, more linguists – and especially more Cognitive Linguists – should start looking at the neurocognitive side of things. More linguists – and especially more Cognitive Linguists – should set out to explore what remains a largely unknown part of God’s truth, instead of exclusively devoting themselves to the mapping of psychological reality. And, crucially, God’s truth linguists and hocus-pocus linguists at large should keep talking to one another, in an effort to inform each other’s ventures into uncharted territory.

4. Conclusion

The relative lack of “neurocognitive depth” in Cognitive Linguistics, on the one hand, and the scarcity of coverage of Cognitive Linguistics in broadly based introductions to cognitive science, on the other hand, provide powerful arguments for a soul-searching exercise. The time has come to take stock, not only of the achievements, but also of the possible vulnerability of Cognitive Linguistics. I consider the scarcity of neurocognitive research within Cognitive Linguistics to be its Achilles’ heel (in the sense that Cognitive Linguistics has not yet engaged in it with sufficient visibility). It is reassuring to see that some Cognitive Linguists, and several onlookers, have been increasingly vocal in this respect and fully endorse the need for an open discussion.

I, for one, have argued that Cognitive Linguistics (as it currently stands) has to broaden its scope even further than it has done hitherto. For now, it is essentially just another competing linguistic model – an attractive one, for sure, but for linguistic-theoretical reasons, not because of an all-encompassing cognitive outlook (one which visibly includes neurocognitive issues). Biological reality is to be taken more seriously. I have pointed out that more Cognitive Linguists (but by no means all of them) will need to follow the lead taken by colleagues such as Lakoff, Deane and Lamb. More of us need to be doing – and to be seen to be doing – the same sort of work.

But perhaps I am missing something. To quote Langacker again:

Are we really doing so badly? I notice that the pages of Cognitive Linguistics are starting to fill up with the results of experimental investigations, and that is symptomatic of what is happening in the field in general.

If that is the case, let us make sure that this new exciting research is given increased visibility. Cognitive Linguists must combat the widespread feeling out there that all they are good at is prototype theory, conceptual metaphor, blending and other such phenomena (i.e. psychological reality). The best way to combat that feeling is by shifting attention to other (neurocognitive) issues, but without neglecting the (analytical) work that has rightly turned Cognitive Linguistics into a force to be reckoned with.30

If, on the other hand, my impression is correct and Langacker’s is premature, then those of us who feel that the connection between Cognitive Linguistics and cognitive science remains weak have a right to speak out. If my impression is correct, the worst that unconvinced Cognitive Linguists can do is put their heads in the sand and hope for the clamours and murmurs to go away. They will not. There has to be a clear recognition that neurocognitive linguistics and analytical cognitive linguistics are both valid forms of Cognitive Linguistics, but that, in the interest of outside recognition, increased visibility, and greater integration with the other cognitive sciences, and to improve the standing of Cognitive Linguistics both among linguists and in the cognitive science community as a whole, more research activity than is currently being undertaken by Cognitive Linguists is needed in the neurocognitive arena. Only then will we be able to truthfully state that Cognitive Linguistics does indeed live up to its name.

References

Achard, Michel. 1998. Representation of Cognitive Structures. Syntax and Semantics of French Sentential Complements. Berlin: Mouton de Gruyter.

Burling, Robbins. 1964. “Cognition and Componential Analysis: God’s Truth or Hocus-Pocus?”. American Anthropologist 66.20-28. Reprinted (1968) in R. A. Manners and D. Kaplan (eds), Theory in Anthropology. A Source-Book. Chicago: Aldine, 514-523.

Casad, Eugene H. (ed.). 1996. Cognitive Linguistics in the Redwoods. The Expansion of a New Paradigm in Linguistics. Berlin: Mouton de Gruyter.

Chomsky, Noam. 1959. Review of Skinner (1957). Language 35.26-58.

Chomsky, Noam. 1979. Language and Responsibility. Sussex: Harvester Press.

Chomsky, Noam. 1995. The Minimalist Program. Cambridge: MIT Press.

Cuyckens, Hubert and Zawada, Britta (eds). 1999. Polysemy in Cognitive Linguistics. Amsterdam: John Benjamins.

Deane, Paul D. 1992. Grammar in Mind and Brain. Explorations in Cognitive Syntax. Berlin: Mouton de Gruyter.

Deane, Paul D. 1996. “Neurological Evidence for a Cognitive Theory of Syntax. Agrammatic Aphasia and the Spatialization of Form Hypothesis”. In Casad (1996: 55-115).

Dennett, Daniel C. 1978. “Skinner Skinned”. In id., Brainstorms. Philosophical Essays on Mind and Psychology. Montgomery: Bradford, 53-70.

Edelman, Gerald M. 1987. Neural Darwinism. The Theory of Neuronal Group Selection. New York: Basic Books.

Edelman, Gerald M. 1992. Bright Air, Brilliant Fire. On the Matter of the Mind. New York: Basic Books.

Elman, Jeffrey L.; Bates, Elizabeth A.; Johnson, Mark H.; Karmiloff-Smith, Annette; Parisi, Domenico; Plunkett, Kim. 1996. Rethinking Innateness. A Connectionist Perspective on Development. Cambridge: MIT Press.

Fauconnier, Gilles. 1995. Mental Spaces. Aspects of Meaning Construction in Natural Language. Cambridge: Cambridge University Press.

Fauconnier, Gilles. 1997. Mappings in Thought and Language. Cambridge: Cambridge University Press.

Fillmore, Charles J. 1977. “Topics in Lexical Semantics”. In R. W. Cole (ed.), Current Issues in Linguistic Theory. Bloomington: Indiana University Press. 76-138.

Fillmore, Charles J. 1982. “Frame Semantics”. In S. Yang (ed.), Linguistics in the Morning Calm. Seoul: Hanshin, 111-137.

Fillmore, Charles J. 1985. “Frames and the Semantics of Understanding”. Quaderni di semantica 6.222-254.

Fodor, Jerry. 1983. The Modularity of Mind. An Essay on Faculty Psychology. Cambridge: MIT Press.

Geeraerts, Dirk. 1995. “Cognitive Linguistics”. In J. Verschueren, J.-O. Östman and J. Blommaert (eds), Handbook of Pragmatics. Amsterdam: John Benjamins, 111-116.

Geiger, Richard A. and Rudzka-Ostyn, Brygida (ed.). 1993. Conceptualizations and Mental Processing in Language. Berlin: Mouton de Gruyter.

Givón, Talmy. 1998. “The Functional Approach to Grammar”. In Tomasello (1998: 41-66).

Grzybek, Peter. 1995. “Foundations of Semiotic Proverb Study”. De Proverbio 1 (http://www.deproverbio.com/DPjournal/DP,1,1,95/GRZYBEK.html). Also, with minor stylistic differences, in Proverbium 4.39-85 (1987).

Harris, Zellig S. 1951. Methods in Structural Linguistics. Chicago: University of Chicago Press.

Hilferty, Joseph. 2000. “Grammar, the Lexicon, and Encyclopedic Knowledge. Is There Such a Thing as Informational Encapsulation?” In Peeters (2000).

Houben, J.E.M. 1993. “Bartrhari en de structuur van het Sanskrit: `God’s truth’ of `hocus-pocus’?”. In J. Goedegebuure (ed.), Nieuwe Wegen in Taal- en Literatuurwetenschap. Handelingen van het Eenenveertigste Filologencongres. Tilburg: Tilburg University Press, 143-158.

Householder, Fred W. Jr. 1952. Review of Harris, Zellig S., Methods in Structural Linguistics. International Journal of American Linguistics 18.260-268.

Krikmann, Arvo A. 1971. “Zur Problematik der Metasprache als Ausdruck der Bedeutungsstreuung der Sprichwörter”. Proverbium 17.624-626.

Kuusi, Matti. 1972. “Towards an International Type-system of Proverbs”. Proverbium 19. 699-736.

Lakoff, George. 1987. Women, Fire, and Dangerous Things. What Categories Reveal about the Mind. Chicago: University of Chicago Press.

Lakoff, George and Johnson, Mark. 1980. Metaphors we Live by. Chicago: University of Chicago Press.

Lakoff, George and Johnson, Mark. 1998. Philosophy in the Flesh. The Embodied Mind and its Challenge to Western Thought. New York: Basic Books.

Lakoff, George and Thompson, Henry. 1975. “Dative Questions in Cognitive Grammar”. In R. Grossman, J. San and T. Vance (eds), Papers from the Parasession on Functionalism. Chicago: Chicago Linguistics Society, 337-350.

Lakoff, George and Turner, Mark. 1989. More than Cool Reason. A Field Guide to Poetic Metaphor. Chicago: University of Chicago Press.

Lamb, Sydney M. 1971. “The Crooked Path of Progress in Cognitive Linguistics”. Georgetown University Monograph Series on Languages and Linguistics 24.99-123. Reprinted (1973) in A. Makkai and D. G. Lockwood (eds), Readings in Stratificational Linguistics. University, Alabama: University of Alabama Press, 12-33.

Lamb, Sydney M. 1999. Pathways of the Brain. The Neurocognitive Basis of Language. Amsterdam: John Benjamins.

Lamb, Sydney M. Forthcoming. “Bi-Directional Processing in Language and Related Cognitive Systems”. In M. Barlow and S. Kemmer (eds), Usage-Based Models of Language. Stanford: CSLI.

Langacker, Ronald W. 1987-1991. Foundations of Cognitive Grammar (2 vol.). Stanford: Stanford University Press.

Langacker, Ronald W. 1990. Concept, Image, and Symbol. The Cognitive Basis of Grammar. Berlin: Mouton de Gruyter.

Langacker, Ronald W. 1998. “Conceptualization, Symbolization, and Grammar”. In Tomasello (1998: 1-39).

Levy, Yonata and Kavé, Gitit. 1999. “Language Breakdown and Linguistic Theory”. Lingua 107.95-143.

Newmeyer, Frederick. 1999. “Bridges between Generative and Cognitive Linguistics”. In L. de Stadler and C. Eyrich (eds), Issues in Cognitive Linguistics. Berlin: Mouton de Gruyter, 3-19.

Peeters, Bert. 1996. “Language and the Mind. On Concepts and Values”. Pragmatics and Cognition 4.139-152.

Peeters, Bert. 1998. “Cognitive Musings”. Word 49.225-237.

Peeters, Bert (ed.). 2000. The Lexicon-Encyclopedia Interface. Oxford: Elsevier.

Regier, Terry. 1996. The Human Semantic Potential. Spatial Language and Constrained Connectionism. Cambridge: MIT Press.

Rosenfield, Israel. 1988. The Invention of Memory. A New View of the Brain. New York: Basic Books.

Rudzka-Ostyn, Brygida (ed.). 1988. Topics in Cognitive Linguistics. Amsterdam: John Benjamins.

Schwarz, Monika. 1992. Einführung in die kognitive Linguistik. Tübingen: Francke.

Skinner, B. F. 1957. Verbal Behavior. New York: Appleton-Century-Crofts.

Talmy, Leonard. 2000. Toward a Cognitive Semantics (2 vol.). Cambridge: MIT Press.

Taylor, John R. 1995. “…”. In R. Dirven and J. Vanparys (eds), Current Approaches to the Lexicon. Frankfurt: Lang, …-… [further details to be supplied]

Tomasello, Michael (ed.). 1998. The New Psychology of Language. Cognitive and Functional Approaches to Language Structure. Mahwah: Lawrence Erlbaum.

Uriagereka, Juan. 1999. Review of Chomsky (1995). Lingua 107.267-273.

Endnotes

1

The “cognitive musings” in this paper are a rethink of those that were published under the author’s name in the August 1998 issue of the journal Word. They owe their existence, and much of their formulation, to René Dirven, who kindly requested that I take part in the “Linguistics and Ideology” theme session at the 6th International Cognitive Linguistics Conference held in Stockholm in July 1999. I hope he does not now regret his insistence, because what I am about to say remains certain to ruffle many feathers. I am grateful to René Dirven, Dirk Geeraerts, Joe Hilferty, Ron Langacker, Sydney Lamb, Francis Steen and John Taylor for having provided me with ideas for which, of course, they do not bear any responsibility. I have also profited from discussions on Funknet, the electronic discussion forum dealing with issues in (American style) functional linguistics based at Rice University (” funknet@listserv.rice.edu “), and on Cogling, Cognitive Linguistics’ own electronic discussion forum based at the University of California San Diego (”cogling@ucsd.edu”).

2

Minimalism (Chomsky 1995), the successor of Government and Binding, itself an outgrowth of classical transformational grammar, does not. Although a lot of excess apparatus has been disposed of, the remnants of previous approaches are too numerous for the framework to be really “minimalist”.

3

Schwarz (1992) did use the label cognitive linguistics to refer to the wider field of “approaches to natural language as a mental phenomenon” (Geeraerts 1995: 112). This includes not only Cognitive Linguistics, but also approaches such as those taken by Noam Chomsky and by Manfred Bierwisch. Schwarz’s example has been followed by, e.g., Taylor (1995) and Newmeyer (1999). Geeraerts (1995) provides a brilliant summary of how Cognitive Linguistics and generative grammar differ in their commitment to cognition.

4

Rohrer (this volume) argues for a broad-based theoretical framework which would tie Cognitive Linguistics in with cognitive science. The latter is redefined as a patient-, problem- and pragmatically-centered multi-disciplinary entreprise that binds together levels of investigation ranging from the cognitive neurosciences through the computer sciences and psychology to anthropology.

5

In fact, as the following quote indicates, he claimed that the terms had been used before. To my knowledge, no earlier source has ever been identified.

6

Note that Householder, in true structuralist fashion, talks about languages and not about language. God’s truth is not that “language has a structure”, a view that most linguists nowadays, unless they believe in universal grammar, would reject out of hand. Cognitive Linguists, for instance, rightly start from conceptual structures, which can be reflected in thousands of different ways in the languages of the world, where they are shaped in part by the building blocks of those languages.

7

In the printed literature, the terms resurface (quite exceptionally for the nineties) in Houben (1993) and in Lamb (forthcoming). In the early seventies, they made a fleeting appearance in proverb studies (Krikmann 1971, Kuusi 1972; cf. Grzybek 1995). In the sixties, anthropologists became familiar with them thanks to Burling (1964).

8

I am almost quoting Householder now, not because I want to be facetious, but because this time there really is an earlier source.

9

Itkonen’s quote shows the effects that political correctness is now having on everyday discourse. Why else was it necessary to qualify those very labels that Householder used with no precautions other than the usual inverted commas?

10

Pursuing God’s truth, as defined here, is different from “taking up God’s perspective, which is impossible” (Mark Johnson apud Hutton, this volume). God’s truth does not necessarily correspond to a God’s eye view of truth, to the truth as it exists prior to any description, to the objective truth about the workings of a language. If it did, it would be unattainable, at least according to the Cognitive Linguistics canon set out in Lakoff (1987) but questioned by Jones (this volume). For reasons known to everyone, Cognitive Linguists will be among the first to recognise that Householder’s terminology is metaphorical. “Because of the pervasiveness of metaphor in thought, we cannot always stick to discussions of reality in purely literal terms” (Lakoff apud Hutton, this volume). This is true as well when we go one step further and actually set out to study the brain (cf. the quote from Lakoff in section 1.2).

11

Until the present day, it has been impossible to track down the exact source of this amazing aphorism. I would love to be in a position to claim authorship for it. As it happens, I encountered it somewhere or other, but I failed to write down who had said or written it first.

12

The team is based at the International Computer Science Institute at Berkeley. There are of course other collaborators whom I have not mentioned. More information is available on the Institute’s website ( http://www.icsi.berkeley.edu/NTL/ ).

13

The entire interview is a worthwhile read for those who are interested in the philosophical revolution that Lakoff and Johnson have been involved in. It can be found in the 51st issue of the electronic magazine Edge ().

14

What matters here is not what (and how much) Lakoff has produced, but what (and how much) he has produced that is actually being read and/or taken further by a majority of Cognitive Linguists.

15

Sego is now the editor of the electronically published Pedagogical Quarterly of Cognitive Linguistics ().

16

Lakoff preferred a more “commercial” title, one that would sell – and sell it did (100,000 copies, according to René Dirven, p.c.).

17

Geiger/Rudzka-Ostyn (1993) is the third volume (after Langacker 1990 and Deane 1992) in the now well-established CLR (Cognitive Linguistics Research) series published, together with the journal Cognitive Linguistics, by Mouton de Gruyter. Among the later volumes, Casad (1996) and Achard (1998) deserve special mention.

18

The upper case initials are in the original text. Work published by the three authors referred to by Lakoff includes Talmy (2000) (a revision and digest of earlier work), Langacker (1987/1991, 1990), Fauconnier (1995, 1997). Talmy introduced the principles of Gestalt psychology into linguistic analysis. Langacker, among other things, developed Talmy’s insights into a coherent overall framework. Fauconnier brought the philosophical questions of reference and mental representation to bear on Cognitive Linguistics.

19

Perhaps neurocognitive linguistics is the answer of one Cognitive Linguist (Lamb) to Givón’s (1998: 64) call for a “combined metadiscipline that is yet to be born cognitive neuro-linguistics”. This is not quite the same as what is commonly referred to as neurolinguistics (tout court). The latter predominantly looks at language disorders (agrammatism, selective language impairments and other aphasias). For a recent “tutorial overview”, see Levy/Kavé (1999).

20

In saying this, I rely on my experience with the cognitive science texts that I have seen over the last few years. I could have produced a list, and in fact tried to as I was writing this paper. But that list soon became unwieldy, even though I had set 1995 as a terminus post quem for my planned survey (which would have required a full-fledged bibliographical report of its own).

21

Chomsky’s review, published about two decades before Dennett (1978), could no doubt be dubbed the “Skinner Skinned” of linguistics.

22

The Cognitive Linguistics heading in Hudson’s bibliography contains items by Cognitive Linguists such as René Dirven, Dirk Geeraerts, Ron Langacker, Günter Radden, John Taylor and Marjolijn Verspoor. The innateness and modularity headings have entries by authors who are at best interested onlookers.

23

Produced for the BBC Horizon series by P. Millson and directed by D. Sington.

24

Does our training as linguists preclude us from understanding what is going on in the brain? Most of us have had no training whatsoever in the workings of the brain. Those who have developed an interest have typically done so independently of their study of linguistics. They have read up on the literature, in a slow but certain process of familiarisation with a hugely complex area which has traditionally been the hunting ground of neurologists, anatomists, brain surgeons and the like.

25

For an early introduction to Edelman, cf. Rosenfield (1988).

26

The assumption, here, is of course that Lamb did not misunderstand Edelman, and that his own hypothesis is correct. A reader of this paper suggested I attempt at least to summarise the basic idea. Suffice it to refer, once more, to the phrase neural Darwinism. Principles applied by Darwin to explain evolution at large apply within the brain as well: strengthening (of the more active neuronal groups), weakening/withering/disappearance (of the less active or inactive ones), in brief “survival of the fittest”.

27

I am grateful to David Tuggy for suggesting, at the 5th International Cognitive Linguistics Conference (Amsterdam, 1997), that for once the appropriate answer to a question which asks for information is affirmative.

28

A theory of language is operationally plausible if it provides a plausible basis for understanding the processes of speaking and comprehension. It is developmentally plausible if there is a plausible means whereby the proposed model or system could be acquired by children. It is neurologically plausible if it can offer a plausible account of how the system might be represented in neural structures. Theories which meet all three criteria qualify as forms of neurocognitive linguistics (in Lamb’s meaning of the term).

29

Fillmore was one of the first to analyse the commercial transaction scene at great length.

30

For an interesting attempt to link prototype semantics and neurocognitive linguistics, cf. Howard (this volume).

Adele E. Goldberg

Summary A new theoretical approach to language has emerged in the past 10-15 years that allows linguistic observations about form-meaning pairings—constructions–to be stated directly. Constructionist approaches aim to account for the full range of facts about language, without assuming that a particular subset of the data is part of a privileged “core”. Researchers argue that unusual constructions shed light on more general issues, and serve to illuminate what is required for a complete account of language.

Keywords: language, linguistics, constructions, grammar, learning, generalizations

Constructions–form and meaning pairings–have been the basis of many major advances in the study of grammar since the days of Aristotle. Observations about particular linguistic constructions have shaped our understanding of both particular languages and the nature of Language itself. But only recently has a new theoretical approach emerged that allows observations about constructions to be stated directly, providing long-standing traditions with a framework that allows both broad generalizations and more limited patterns to be analyzed and accounted for fully. Many linguists with varying backgrounds have converged on several key insights that have given rise to a family of approaches, here referred to as constructionist approaches [2-23].

Constructionist approaches share certain foundational ideas with the mainstream “generative” approach that has held sway for the past several decades[1,36,37]. Both approaches agree that it is essential to consider language as a cognitive (mental) system; both approaches acknowledge that there must be a way to combine structures to create novel utterances, and both approaches recognize that a non-trivial theory of language learning is needed.

In other ways, constructionist approaches contrast sharply with the mainstream generative approach. The latter has held that the nature of language can best be revealed by studying formal structures independently of their semantic or discourse functions. Ever increasing layers of abstractness have characterized the formal representations. Meaning is claimed to derive from the mental dictionary of words, with functional differences between formal patterns being largely ignored. Semi-regular patterns and cross-linguistically unusual patterns are viewed as “peripheral,” with a narrowing band of data seen as relevant to the “core” of language. Mainstream generative theory argues further that the complexity of core language cannot be learned inductively by general cognitive mechanisms and therefore learners must be hard-wired with principles that are specific to language (“universal grammar”).

Each basic tenet outlined below is shared by most constructionist approaches. Each represents a major divergence from the mainstream generative approach, and a return in many ways to a more traditional view of language.

1) All levels of description are understood to involve pairings of form with semantic or discourse function, including morphemes or words, idioms, partially lexically filled and fully abstract phrasal patterns. (See Box 1)

2) An emphasis is placed on subtle aspects of the way we conceive of events and states of affairs.

3) A “what you see is what you get” approach to syntactic form is adopted: no underlying levels of syntax, nor any phonologically empty elements are posited.

4) Constructions are understood to be learned on the basis of the input and general cognitive mechanisms (they are constructed), and are expected to vary cross linguistically.

5) Cross-linguistic generalizations are explained by appeal to general cognitive constraints together with the functions of the constructions involved.

6) Language-specific generalizations across constructions are captured via inheritance networks much like those that have long been posited to capture our non-linguistic knowledge.

7) The totality of our knowledge of language is captured by a network of constructions: a “construct-i-con.”

Each of these tenets is explained in a subsequent section below.

Constructions: what they are

Constructions are stored pairings of form and function, including morphemes, words, idioms, partially lexically filled and fully general linguistic patterns. Examples are given in Box 1.

Morphemee.g., anti-, pre-, -ing

Worde.g., Avocado, anaconda, and

Complex worde.g., Daredevil, shoo-in

Idiom (filled)e.g., Going great guns

Idiom (partially filled)e.g., Jog memory

Covariational Conditional

construction [10]Form: The Xer the Yer

(e.g., The more you think about it,

the less you understand) Meaning: linked independent

and dependent variables; see text.

Ditransitive (double object) constructionForm: Subj [V Obj1 Obj2]

(e.g., He gave her a Coke;

He baked her a muffin.) Meaning: transfer (intended or actual); see text.

PassiveForm: Subj aux VPpp (PPby)

(e.g., The armadillo was hit

by a car)Discourse function: to make undergoer topical and/or actor

non-topical

Box 1. Examples of constructions, varying in size and complexity; form and function are specified if not readily transparent.

Any linguistic pattern is recognized as a construction as long as some aspect of its form or function is not strictly predictable from its component parts or from other constructions recognized to exist. In addition, many constructionist approaches argue that patterns are stored even if they are fully predictable as long as they occur with sufficient frequency [24-29].

Unlike mainstream generative grammar, the framework emphasizes the semantics and distribution of particular words, grammatical morphemes, and cross-linguistically unusual phrasal patterns; the hypothesis behind this methodology is that an account of the rich semantic/pragmatic and complex formal constraints on these patterns readily extends to more general, simple or regular patterns.

As an example of an unusual pattern, consider the Covariational Conditional construction in Box 1 (e.g., The more you think about it, the less you understand). The construction is interpreted as involving an independent variable (identified by the first phrase) and a dependent variable (identified by the second phrase). The word the normally occurs at the beginning of a phrase headed by a noun. But in this construction it requires a comparative phrase. The two major phrases of the construction resist classification as either noun phrases or clauses. The requirement that two phrases of this type be juxtaposed without conjunction is another non-predictable aspect of the pattern. Because the pattern is not strictly predictable, a construction is posited that specifies the particular form and semantic function involved [10].

Other unusual constructions, with example instances to the right, include those in Box 2. While each pattern may be primarily colloquial, it is part of every native speaker’s repertoire of English. (The stranded preposition construction is unusual not in that it is prescriptively dispreferred, but that it is only found in a few Germanic languages).

time away constructionTwistin the night away[13]

What’s X doing Y?What’s that fly doing in my soup?[30]

Nominal Extraposition constructionIt’s amazing the difference![31]

Mad Magazine construction

Him, a doctor?![32]

N P N construction

house by house; day after day[12]

Stranded preposition constructionWho did he give that to?

Box 2: Productive or semi-productive constructions that are unusual cross-linguistically and must be learned on the basis of the input.

More common patterns such as passive, topicalization and relative clauses are understood to be learned pairings of form and (semantic or discourse) function—constructions, as well. Each pairs certain formal properties with a certain communicative function.

Even basic sentence patterns of a language can be understood to involve constructions. That is, the main verb can be understood to combine with an argument structure construction (e.g., transitive, intransitive, ditransitive constructions etc.) [7]. The alternative is to assume that the form and general interpretation of basic sentence patterns of a language are determined by semantic and/or syntactic information specified by the main verb. The sentence patterns given in (1) and (2) indeed appear to be determined by the specifications of give and put respectively:

1. Chris gave Pat a ball.

2. Pat put the ball on the table.

Give is a three argument verb. An act of giving requires three characters: a giver (or agent), a recipient, and something given (or “theme”). It is therefore expected to appear with three phrases corresponding to these three roles. In (1), for instances, Chris is agent, Pat is recipient, and a ball is theme. Put, another three argument verb, requires an agent, a theme (object that undergoes the change of location) and a final location of the theme’s motion. It appears with the corresponding three arguments in (2). However, while (1) and (2) represent perhaps the prototypical case, in general the interpretation and form of sentence patterns of a language are not reliably determined by independent specifications of the main verb. For example, it is implausible to claim that sneeze has a three argument sense, and yet it can appear in (3). The patterns in (4)-(6) are likewise not naturally attributed to the main verbs:

3. “He sneezed his tooth right across town.” (Andrew’s Loose Tooth, Robert Munsch)

4. “She smiled herself an upgrade.” (A. Douglas, Hitchhiker’s guide to the Galaxy Harmony Books)

5. ‘We laughed our conversation to an end.’ (J. Hart. 1992, Sin NY: Ivy Books)

6. “They could easily co-pay a family to death.” (NYT, 1/14/02)

Examples need not be particularly novel to make the point. Verbs typically appear with a wide array of complement configurations. Consider the verb slice and the various constructions in which it can appear (labeled in parentheses):

7a. He sliced the bread. (transitive)

b. Pat sliced the carrots into the salad. (caused motion)

c. Pat sliced Chris a piece of pie. (ditransitive)

d. Emeril sliced and diced his way to stardom. (way construction)

e. Pat sliced the box open. (resultative)

In all of these expressions slice means to cut with a sharp instrument. It is the argument structure constructions that provide the direct link between surface form and general aspects of the interpretation such as something acting on something else (7a), something causing something else to move(7b), someone intending to cause someone to receive something (7c), someone moving somewhere(7d), someone causing something to change state(7e)[7,33].

Thus constructions can be seen to be essential to an effective account of both unusual or especially complex patterns and for the basic, regular patterns of language.

The functions of constructions

Different surface forms are typically associated with slightly different semantic or discourse functions. Take for example, the ditransitive construction, which involves the form, Subj V Obj1 Obj2 (e.g., (1), (8b), (9b)). The ditransitive form evokes the notion of transfer or “giving.” This is in contrast to possible paraphrases. For example, while (8a) can be used to mean that Liza bought a book for a third party because Zach was too busy to buy it himself, (8b) can only mean that Liza intended to give Zach the book. Similarly while (9a) can be used to entail caused motion to a location(the book is caused to go to storage), the ditransitive pattern requires that the goal argument be an animate being, capable of receiving the transferred item (cf.9b, 9c). As is clear from considering the paraphrases, the implication of transfer is not an independent fact about the words involved. Rather the implication of transfer comes from the ditransitive construction itself.

(8) a. Liza bought a book for Zach.

b Liza bought Zach a book.

(9) a. Liza sent a book to storage.

b. Liza sent Stan a book.

c. c. ??Liza sent storage a book.

Other interpretations for the ditransitive can also be systematically related to the notion of transfer, in that they may imply that the transfer will occur if certain satisfaction conditions evoked by the main verb occur (10a), that transfer will not occur (10b), or that the antonymic relation of giving, that of taking away occurs (10c). Even examples such as Cry me a river can be related to the notion of giving via a metaphorical extension[7].

(10) a. Liza guaranteed Zach a book. (If the guarantee is satisfied, Z. will receive a book)

b. Liza refused Zach a book. (Liza caused Zach not to receive a book)

c. Liza cost Zach his job. (Liza causes Zach to lose his job).

In addition to semantic generalizations there also exist generalizations about information structure properties of the construction, or the way in which a speaker’s assumptions about the hearer’s state of knowledge and consciousness at the time of speaking is reflected in surface form. In particular, there is a reliable statistically tendency for the recipient argument to have already been mentioned in the discourse (often encoded by a pronoun) as compared to prepositional paraphrases [9,34,35]. Facts about the use of entire constructions, including register (e.g. formal or informal), dialect variation, etc. are stated as part of the construction as well. Constructionist approaches provide a direct way of accounting for these facts, since constructions specify a surface form and a corresponding function.

The form of constructions

In order to capture differences in meaning or discourse properties between surface forms, constructionist theories do not derive one construction from another, as is typically done in mainstream generative theory. An actual expression or construct typically involves the combination of at least half a dozen different constructions. For example, the construct in (11) involves the list of constructions given in (12a-f):

[color coded]

(11) [ What did Liza buy the child? ]

(12) a. Liza, buy, the, child, what, did constructions (i.e. words)

b. Ditransitive construction

c. Question construction

d. Subject-Auxiliary inversion construction

e. VP construction

f. NP construction

Note that “surface form” need not specify a particular word order, nor even particular grammatical categories, although there are constructions that do specify these features. For example, the ditransitive construction in (11) and discussed above is characterized in terms of a set of argument types. The overt order of arguments in (11) is determined by a combination of a verb phrase construction with the Question construction, the latter of which allows for the “theme” argument (represented by What) to appear sentence initially.

Constructions are combined freely to form actual expressions as long as they are not in conflict. For example, the specification of the ditransitive construction that requires an animate recipient argument conflicts with the meaning of storage in (9c) resulting in unacceptability. The observation that language has an infinitely creative potential [1, 36] is accounted for, then, by the free combination of constructions.

Learning constructions

The fourth tenet states that constructions are understood to be learned on the basis of positive input and to vary cross linguistically. This idea highlights a major difference between most constructional approaches and most mainstream generative approaches, since the latter have argued that learners must be hard-wired with principles specific to a language faculty or “universal grammar” [37, see also 21].

Crucially, all linguists recognize that a wide range of semi-idiosyncratic constructions exist in every language, constructions that cannot be accounted for by general, universal or innate principles or constraints (e.g., examples in Box 2). Mainstream generative theory has taken the position that these constructions exist only on the “periphery” of language—that they need not be the focus of linguistic or learning theorists.[37] Constructionist approaches on the other hand have zeroed in on these constructions, arguing that whatever means we use to learn these patterns can easily be extended to account for so-called “core” phenomena. In fact, by definition, the core phenomena are more regular, and tend to occur more frequently within a given language as well. Therefore if anything, they are likely to be easier to learn. Since every linguist would presumably agree that the “peripheral,” difficult cases must be learned inductively on the basis of the input, constructionist theories propose that there is no reason to assume that the more general, regular, frequent cases cannot possibly be.

In fact, constructionist theories argue that language must be learnable from positive input together with fairly general cognitive abilities[18,29,38], since the diversity and complexity witnessed does not yield to accounts that assume that cross-linguistic variation can be characterized in terms of a finite set of parameters [37, 57]. Research in this area is quickly gaining momentum. A number of constructionists have made good on the promise to explain how particular constructions are learned [26,27]. It turns out that the input may not be nearly as impoverished as is sometimes assumed [39]; analogical processes can be seen to be viable once function as well as form is taken into account[40,41]; there is good reason to think that children’s early grammar is quite conservative, with generalizations emerging only slowly [29,42,43]; and the ability to record transitional probabilities and statistical generalizations in the input has proven a powerful means by which to learn certain types of generalizations [44].

This approach takes a somewhat different view of what is universal about language than mainstream generative theory. Linguists generally talk of certain constructions as existing in many languages, e.g., the passive construction, relative clause construction, question construction, etc. However, two constructions in different languages can be identified as instances of the same construction if and only if their form and function is identical once other constructions in the language that may differ are factored out. In point of fact, this rarely occurs except in cases of shared diachronic history or language contact [20,45,46]. What is truly remarkable is the degree to which human languages differ from one another, given that all languages need to express roughly the same types of messages. Constructionist approaches anticipate such fairly wide variability across languages [47,48].

Reference to the “same” construction in unrelated languages can be made sense of by understanding that what is intended by such references are actually types of constructions. Two constructions may be, for example, of the passive type in that they share certain functional and formal characteristics even if they are not identical. That is, two constructions in different languages can be identified as instances of the same type of construction if and only if they serve a closely related function and form.

Cross-linguistic generalizations

A driving question behind much of linguistic research is, what is the typology of possible constructions and what constrains it? Constructionist approaches often turn to grammar-external explanations such as universal functional pressures, iconic principles, and processing and learning constraints to explain such empirically observable cross-linguistic generalizations. For example, certain generalizations about how form and meaning tend to be linked cross-linguistically can be explained by appeal to iconic and analogical processes [6,35,49-51,55]. Constraints on long-distance dependency constructions (traditional “island constraints”) appear to yield to processing explanations that take into account the function of the constructions involved [19, 52-54]. Processing accounts have also been suggested to account for certain alternative word order options [55,56].

Even among generative linguists there has been a trend toward the view that many constraints on language that have traditionally been seen as requiring recourse to innate stipulations that are specific to language can actually be explained by general cognitive mechanisms. For example, the fact that that all languages seem to have noun and verb (and possibly adjective) categories may be explained by the existence of corresponding basic semantic categories [57]. Hauser, Chomsky and Fitch go so far as to suggest that the only language-specific innate ability that may be required is recursion, and they raise the point that even that may turn out not to be specific to language [58].

Intra-language generalizations

Inheritance hierarchies have long been found useful for representing all types of knowledge, e.g., our knowledge of concepts. The construction-based framework captures linguistic generalizations within a particular language via the same type of inheritance hierarchies [2, 59-60]. Broad generalizations are captured by constructions that are inherited by many other constructions; more limited patterns are captured by positing constructions that are at various midpoints of the hierarchical network. Exceptional patterns are captured by low level constructions. For example, the “What’s doing ?” construction, which has a fixed form and connotes some sort of unexpectedness, captures a pattern in the grammar of English. It inherits properties from several other more general constructions, including the Left Isolation, the Subject Auxiliary Inversion, the Subject-Predicate and the Verb-Phrase constructions [30].

Constructions all the way down

What makes a theory that allows for constructions a “construction-based” theory is tenet 7: the idea that the network of constructions captures our knowledge of language in toto i.e., it’s constructions all the way down.

Conclusion

Constructionist theories set out to account for all of our knowledge of language as patterns of form and function. That is, the constructionist approach does not assume that language should be divided up into “core” grammar and the to-be-ignored “periphery.” In identifying constructions, an emphasis is placed on subtle aspects of construal and on surface form. Cross-linguistic generalizations are explained by appeal to general cognitive constraints together with the functions of the constructions involved. Language-specific generalizations across constructions are captured via inheritance networks. The inventory of constructions, which includes morphemes or words, idioms, partially lexically filled and fully abstract phrasal patterns, is understood to be learned on the basis of the input together with general cognitive mechanisms.

Major questions

Do there exist generalizations about form that do not have even an abstract, family-resemblance or radial category type generalization about function associated with them?

How does the full range of phenomena considered by mainstream generative grammarians translate into a constructional approach?

Acknowledgements

I am grateful to Ray Jackendoff, Fritz Newmeyer, Devin Casenhiser, Mike Tomasello, Ali Yazdani and three anonymous referees for advice on an earlier draft. I am sure I will regret any I failed to heed.

References

1 Noam Chomsky, (1957) Syntactic Structures. The Hague: Mouton.

2 George Lakoff, (1987) Women, fire, and dangerous things: what categories reveal about the mind Chicago: University of Chicago Press.

3 Ronald W. Langacker, (1987/1991) Foundations of cognitive grammar Volume I & II Stanford, Calif.: Stanford University Press.

4 Charles J. Fillmore et al. (1988) Regularity and Idiomaticity in Grammatical Constructions: The Case of Let Alone, Language 64: 501-538.

5 Anna Wierzbicka, (1988) The Semantics of Grammar Amsterdam: John Benjamins Publishing.

6 Knud Lambrecht, (1994) Information Structure and Sentence Form Cambridge: Cambridge University Press.

7 Adele E. Goldberg, (1995) Constructions: A Construction Grammar Approach to Argument Structure Chicago: Chicago University Press.

8 L. Gleitman et al. (1996), ‘Similar’ and similar concepts, Cognition 58, 321-76. .

9 Sandra A. Thompson, (1990) Information Flow and Dative Shift in English Discourse, in Development and Diversity: Linguistic Variation across Time and Space., ed. Jerrold Edmondson, Katherine Feagin, and Peter MühlhäuslerSummer Institute of Linguistics, 239-253.

10 Culicover, Peter W. and Ray Jackendoff. (1999) The view from the periphery: the English comparative correlative. Linguistic Inquiry 30 (4): 543-571.

11 Arnold Zwicky. (1994) Dealing out meaning: Fundamentals of Syntactic Constructions, Berkeley Linguistics Society 20: 611-625.

12 Edwin Williams. (1994) Remarks on Lexical Knowledge, Lingua 92: 7-34.

13 Ray Jackendoff. (1997) Twistin’ the night away, Language 73, no. 3: 534-559.

14 Ivan A Sag. (1997) English Relative Clause Constructions, Journal of Linguistics 33, no. 2: 431-484.

15 Gert Webelhuth and Farrell Ackerman, (1998) A Theory of Predicates Stanford: CSLI Publications. Distributed by Cambridge University Press.

16 Seizi Iwata, (1998) A Lexical Network Approach to Verbal Semantics Tokyo: Kaitakusha.

17 Masayoshi Shibatani. (1999) Dative subject constructions 22 years later, Studies in the Linguistic Sciences 29, no. 2: 45-76.

18 Peter W. Culicover, (1999) Syntactic Nuts: Hard Cases in Syntax Oxford: Oxford University Press.

19 Robert Van Valin Jr., (1998) `The acquisition of WH-questions and the mechanisms of language acquisition’, M. Tomasello, ed., The New Psychology of Language: Cognitive and Functional Approaches to Language Structure, 221-49. Hillsdale, N.J.: LEA.

20 William Croft, (2001) Radical Construction Grammar Oxford: Oxford University Press.

21 Ray Jackendoff, (2002) Foundations of Language Oxford: Oxford University Press.

22 Bybee, Joan. (2001) Main clauses are innovative, subordinate clauses are conservative: consequences for the nature of constructions. In J. Bybee and M. Noonan (eds.) Complex sentences in grammar and discourse: essays in honor of Sandra A. Thompson. Amsterdam: John Benjamins.

23 Geert Booij. (2002) Constructional Idioms, Morphology, and the Dutch Lexicon. Journal of Germanic Linguistics 14 4:301-329

24 Ronald W. Langacker, (1988) A usage-based model., in Topics in cognitive linguistics, ed. B. Rudzka-Ostyn Philadelphia: John Benjamins. 127-161

25 Michael Barlow and Suzanne Kemmer, (2000) Usage Based Models of Grammar Stanford: CSLI Publications.

26 Michael Israel et al. (2000) From States to Events: the Acquisition of English Passive Participles, Cognitive Linguistics 11, no. 1/2: 1-27.

27 Holger Diessel and Michael Tomasello. (2001) The acquisition of finite complement clauses in English: A usage based approach to the development of grammatical constructions., Cognitive Linguistics 12: 97-141.

28 Arie Verhagen (2002) From Parts to Wholes and Back Again. Cognitive Linguistics 13-4

29 Michael Tomasello, (in press) Constructing a language: A Usage-Based Theory of Language Acquisition Harvard University Press.

30 Paul Kay and Charles J. Fillmore. (1999) Grammatical constructions and linguistic generalizations: The What’s X doing Y? construction, Language 75, no. 1: 1-34.

31 Laura A. Michaelis and Knud Lambrecht. (1996) Toward a Construction-Based Model of Language Function: The Case of Nominal Extraposition, Language 72: 215-247.

32 Knud Lambrecht. (1990) “What, me worry?” Mad Magazine sentences revisited, Proceedings of the 16th Annual Meeting of the Berkeley Linguistics Society. Berkeley, California. : 215-228.

33 Adele E. Goldberg, (in press) Argument Realization: the role of constructions, lexical semantics and discourse factors., in Construction Grammar(s): Cognitive and Cross-language dimensions, eds. Mirjam Fried and Jan-Ola Östman. John Benjamins.

34 N. Erteschik-Shir, (1979) Discourse Constraints on Dative Movement, in Syntax and Semantics, ed. Suzanne Laberge and Gillian Sankoff New York: Academic Press, 441-467.

35 Thomas Wasow. (2002) Postverbal Behavior. Stanford: CSLI Publications.

36 Noam Chomsky (1965) Aspects of the Theory of Syntax. Cambridge, Mass: MIT Press.

37 Noam Chomsky. (1981) Lectures on Government and Binding. Foris, Dordrecht.

38 Jeffrey Elman et al. (1996) Rethinking Innateness: A Connectionist Perspective on Development Cambridge, Mass: MIT Press.

39 Geoffrey K. Pullum and Barbara C. Scholz. (2002) Empirical assessment of stimulus poverty arguments, The Linguistic Review 19, no. 1-2: 9-50.

40 Adele E. Goldberg, (1999) The Emergence of Argument Structure Semantics, in The Emergence of Language, ed. B. MacWhinney Lawrence Erlbaum Publications.

41 Michael Israel. (2002) Consistency and Creativity in First Language Acquisition, Proceedings of the Berkeley Linguistic Society 29.

42 Elena V. M. Lieven et al. (1997) Lexically-based learning and early grammatical development, Journal of Child Language 24, no. 1: 187-219.

43 Michael Tomasello. (2000) Do Young Children Have Adult Syntactic Competence?, Cognition 74 (3): 209-253

44 Jenny R. Saffran. (2001) The use of predictive dependencies in language learning, Journal of Memory and Language 44: 493-515.

45 Betty Birner and Gregory Ward, (1998) Information Status and Noncanonical Word Order in English Philadelphia: John Benjamins.

46 Ning Zhang. (1998) The interactions between construction meaning and lexical meaning, Linguistics 36, no. 5: 957-980.

47 William A. Foley and Robert Van Valin Jr., (1984) Functional Syntax and Universal Grammar Cambridge: Cambridge University Press.

48 J. Garry and C. Rubino, eds. Facts about the World’s Languages: An Encyclopedia of the world’s major languags past and present (New York: H.W. Wilson, 2001).

49 John Haiman, (1985) Iconicity in Syntax Cambridge: Cambridge University Press.

50 Talmy Givón. (1991) Isomorphism in the Grammatical Code: Cognitive and Biological Considerations, Studies in Language 1, no. 15: 85-114.

51 Suzanne Kemmer and Arie Verhagen, (2002) The grammar of causatives and the conceptual structure of events, in Mouton Classics: from syntax to cognition, from phonology to text Berlin: Mouton de Gruyter, 451-491.

52 Robert Kluender. (1998) On the distinction between strong and weak islands: A processing perspective, Syntax Semantics 29: 241-279.

53 Robert Kluender and Marta Kutas. (1993) Subjacency as a processing phenomenon, Language and Cognitive Processes 8, no. 4: 573-633.

54 Nomi Erteschik-Shir, (1998) The Syntax-Focus Structure Interface, in Syntax and Semantics 29:The Limits of Syntax, ed. P. Culicover and L. McNally, 211-240.

55 J. Hawkins, (1994) A performance theory of order and constituency Cambridge: Cambridge University Press.

56 Haruko Yamashita and Franklin Chang. (2001) “Long before short” preference in the production of a head-final language, Cognition 81, no. 2: B45-B55.

57 Mark Baker, (in press) Verbs, Nouns, and Adjectives: Their Universal Grammar Cambridge: Cambridge University Press.

58 Mark D. Hauser et al. (2002) The Faculty of Language: What is it, who has it, and how did it evolve?, Science .298 (5598): 1569-1579

59 Carl Jesse Pollard and Ivan Sag, (1994) Head-driven phrase structure grammar. Stanford: CSLI Publications.

60 Adele E. Goldberg (in press) Words by Default: Inheritance and the Persian complex Predicate Construction. In Mismatch: Form-Function Inconcruity and the Architecture of Grammar eds. Elaine Francis and Laura Michaelis. Stanford: CSLI Publications.

句式：语言研究的新方法

Gilles Fauconnier

I. MEANING, LANGUAGE, COGNITION

Linguists agree on one thing – that language is diabolically hard to study. They do not always agree, however, on the how’s, the why’s, and the what for’s: how one should go about studying it and how speakers manage to do what they do; why it is so hard and why exactly we bother to study it; what language is for, and what linguistics is for. A mainstream view that has been popular in the last thirty years (but not necessarily before that) offers the following answers.

How linguists do it: they collect grammaticality judgments from natives and concurrently build and check hypotheses about the formal structure of particular languages and languages in general. How humans do it: they come equipped biologically with innate language-specific universals, that require only minimal fine-tuning when exposed to a particular specimen. Why it’s hard: easy for the child who has the innate universals already set up, hard for the linguist lost in a forest of idiosyncrasies that hide the deeper principles. Why bother? So that we can discover such principles.

What is language for? The story here is that this question is not a priority for the scientist. We can worry later about function, communication, and meaning generally. And what is linguistics for? Well, there is the platonic reward of discovering structure for the sake of structure itself. And then there is biology: Since the universals are in the brain, they must also be in the genes; linguistics is theoretical biology; geneticists and neuroscientists will fill in the messy details of its implementation in our bodies.

This strange and simple story contains its own methods and generalizations. The appropriate methods are in the ‘how to do it’ – collecting grammaticality judgments and so on. What counts as generalizations are the formal principles that apply to wider ranges of phenomena and/or languages.

In contrast to this sharply autonomous view of language structure, cognitive linguistics has resurrected an older tradition. In that tradition, language is in the service of constructing and communicating meaning, and it is for the linguist and cognitive scientist a window into the mind. Seeing through that window, however, is not obvious. Deep features of our thinking, cognitive processes, and social communication need to be brought in, correlated, and associated with their linguistic manifestations.

The cognitive linguistics enterprise, we believe, has already been remarkably successful. It is not far-fetched to say that perhaps for the first time a genuine science of meaning construction and its dynamics has been launched. This has been achieved by intensively studying and modeling the cognition that lies behind language and goes far beyond it, but which language reflects in certain ways, and which in turn supports the dynamics of language use, language change, and language organization. Echoing Erving Goffman, I have called this backstage cognition. Language is only the tip of a spectacular cognitive iceberg, and when we engage in any language activity, be it mundane or artistically creative, we draw unconsciously on vast cognitive resources, call up innumerable models and frames, set up multiple connections, coordinate large arrays of information, and engage in creative mappings, transfers, and elaborations. This is what language is about and what language is for. Backstage cognition includes viewpoints and reference points, figure-ground / profile-base / landmark-trajector organization, metaphorical, analogical, and other mappings, idealized models, framing, construal, mental spaces, counterpart connections, roles, prototypes, metonymy, polysemy, conceptual blending, fictive motion, force dynamics.

Well, where does all this come from? Did it all just spring up in the fertile mind of cognitive linguists, giving them an unlimited supply of new notions to draw from in order to explain some linguistic facts that they wish to talk about? And if so, isn’t all this a considerable weakening of linguistic theory, letting in so many flaky new gimmicks that virtually anything at all becomes easily but vacuously explainable?

Mais pas du tout. Rather remarkably, all the aspects of backstage cognition just alluded to receive ample justification on non-linguistic grounds from a variety of sources. Some have been extensively studied in psychology (e.g. prototypes, figure- ground, analogy), others in artificial intelligence and/or sociology (frames, roles, cultural models), literature and philosophy (metaphor). Metonymy, mental spaces, force dynamics, conceptual blending, initially studied primarily by linguists have been shown to apply to cognition generally. The notion of viewpoint and reference point is presumably even more general, given the nature of our visual systems and orientation. Needless to say, all these features of backstage cognition deserve to be studied and understood in their own right, not just as a means of explaining linguistic distributions. To cognitive scientists who are not linguists, the linguistic distributions matter very little. And for cognitive linguists, there has been a major shift of interest. The cognitive constructs, operations, and dynamics, and the understanding of conceptual systems have become a central focus of analysis. The linguistic distributions are just one of many sources of relevant data.

This shift bears on the methods employed and the generalizations obtained. Methods must extend to contextual aspects of language use and to non-linguistic cognition. This means studying full discourse, language in context, inferences actually drawn by participants in an exchange, applicable frames, implicit assumptions and construal, to name just a few. It means being on the look-out for manifestations of conceptual thought in everyday life, movies, literature, and science. This is because introspection and intuition are woefully insufficient to tell us about general operations of meaning construction. When we volunteer a meaning for an isolated sentence, we do it typically on the basis of defaults and prototypes. It is only in rich contexts that we see the full force of creative on-line meaning construction.

As for generalizations, the most powerful ones are those which transcend specific cognitive domains. In our work on conceptual blending, we see as a strong generalization the discovery that the same principles apply to framing, metaphor, action and design, and grammatical constructions. This is not an internal generalization about language, it is an external one relating linguistic phenomena to non-linguistic ones. Such generalizations seem primordial to the understanding of how language relates to general cognition, but they are precluded in principle by the autonomous approach evoked above. It is no surprise, then, if that approach finds no connection between language and the rest of cognition, for that autonomy is built into the very method that serves to build up the field of inquiry and the theories that are its by-products.

Although cognitive linguistics espouses the age-old view that language is in the service of meaning, its methods and results have been quite novel. The results in fact have been somewhat surprising. At the most general level, here are three that I find striking. I will call them respectively Economy, Operational Uniformity, Cognitive Generalization.

ECONOMY AND THE ELIZA EFFECT

By Economy, I mean the following: any language form in context has the potential to trigger massive cognitive constructions, including analogical mappings, mental space connections, reference point organization, blends, and simulation of complex scenes. When we try to spell out backstage cognition in detail, we are struck by the contrast between the extreme brevity of the linguistic form and the spectacular wealth of the corresponding meaning construction. Very sparse grammar guides us along the same rich mental paths, by prompting us to perform complex cognitive operations. What is remarkable is that by and large subjects engage in quite similar constructions on the basis of similar grammatical prompts, and thereby achieve a high degree of effective communication. The reason seems to be that the cultural, contextual, and cognitive substrate on which the language forms operate is sufficiently uniform across interlocutors to allow for a reasonable degree of consistency in the unfolding of the prompted meaning constructions. How this works remains in many ways mysterious. What is clear is that language is radically different from an information carrying and information preserving system, such as a code or telecommunications. Language forms carry very little information per se, but can latch on to rich preexistent networks in the subjects’ brains and trigger massive sequential and parallel activations. Those activated networks are of course themselves in the appropriate state by virtue of general organization due to cognition and culture, and local organization due to physical and mental context. Crucially, we have no awareness of this amazing chain of cognitive events that takes place as we talk and listen, except for the external manifestation of language (sounds, words, sentences) and the internal manifestation of meaning: with lightning speed, we experience meaning. This is very similar to perception, which is also instantaneous and immediate with no awareness of the extraordinarily complex intervening neural events.

What we are conscious of determines our folk-theories of what is going on. In the case of perception, the folk theory, an extremely useful one for us as living organisms, is that everything we perceive is indeed directly the very essence of the object perceived, out there in the world and independent of us. The effect is contained entirely in the cause. In the same way, our folk theory of language is that the meanings are contained directly in the words and their combinations, since that is all that we are ever consciously aware of. The effect (meaning) is attributed essentially to the visible cause (language). And again, this folk-theory is extremely useful to us as human organisms in everyday life. It makes sense. At another level, the level of scientific inquiry, this folk-theory, like other folk-theories, is wrong, and the information processing model of language breaks down. This reveals that, as humans experiencing language, we are fooled by an interesting variant of the Eliza effect. The famous computer program Eliza produced what looked like a sensible interaction between a psychiatrist and a subject operating the program, but the rich meaning that seemed to emanate from the machine was in fact read in (constructed) by the subject. And strikingly, just like a perceptual illusion, this effect cannot easily be suspended by rational denial. In the case of Eliza, the illusion may be hard to block, but it is easy to see. The more general illusion that meaning is in the language forms is both hard to repress and hard to acknowledge. And for that reason, it has made its way into many scientific accounts of language. In such accounts, the notion that forms have meaning is unproblematic, and the “only” problem becomes to give a formal characterization of such meanings associated with forms. Clearly, if the presupposition that there are such meanings is in error, the very foundations of such accounts are in jeopardy. It has been, I believe, a major contribution of cognitive linguistics to dispel this very strong unquestioned assumption.

OPERATIONAL UNIFORMITY

It is commonly thought that very different operations apply to the various levels of linguistic analysis. For example, syntax governs the sentence, and semantics provides it compositionally with a meaning. At a higher level, other quite different operations apply to produce implicatures, derived meaning, indirect speech acts. Then rhetorical and figurative devices may kick in, such as metaphor and metonymy. Our findings suggest a very different picture. Backstage cognition operates in many ways uniformly at all levels. Figure- ground and viewpoint organization pervades the sentence (Talmy (1978).; Langacker (19987/1991), the Tense system (Cutrer (1994)., Narrative structure (Sanders and Redeker (1996)., in signed and spoken languages, and of course many aspects of non- linguistic cognition. Metaphor builds up meaning all the way from the most basic levels to the most sophisticated and creative ones (Lakoff and Turner (1989); Grady (1997)). And the same goes for metonymic pragmatic functions (Nunberg (1978)) and mental space connections (Sweetser and Fauconnier (1996), Van Hoek (1996), Liddell (1996), which are governed by the same general Access principle. Frames, schemas and prototypes account for word level and sentence level syntactic/semantic properties in cognitive and construction grammar (Lakoff (1987), Fillmore (1985), Goldberg (1997), Langacker (1987/91)), and of course they guide thought and action more generally (Bateson (1972), Goffman (1974), Rosch;). Conceptual blending and analogy play a key role in syntax and morphology (Mandelblit (1997)), in word and sentence level semantics (Sweetser), and at higher levels of reasoning and rhetoric (Robert (1998), Coulson (1997), Turner (1996) ). Similarly, we find force dynamics and fictive motion (Talmy (1985, 1998) operating at all levels (single words, entire systems, like the modals, and general framing).

This operational uniformity is unexpected, remarkable, and counter-intuitive. It has taken cognitive linguists a lot of hard work and theoretical conceptual rethinking to uncover this series of powerful generalizations. There are quite a few interesting reasons for the difficulty of thinking in this new way. One is that language does not come with its backstage cognition neatly displayed ‘on its sleeve’. Everything that counts is deeply hidden from our consciousness, and masked by the ‘folk theory’ effects mentioned earlier. Another difficulty has to do with the long tradition of apprehending limited aspects of language in a self- contained, language-specific, descriptive apparatus. The resulting specialized technical vocabulary has been immensely helpful in launching a coherent linguistic science, but regrettably it has also shielded linguistics from a more comprehensive cognitive framework in which the right questions could be asked.

COGNITIVE GENERALIZATION

Operational uniformity, as outlined in the previous section, pertains essentially to language and reasoning. The uniformity is across linguistic levels, the word, the sentence, the sentence and its context, the whole discourse, and ultimately general reasoning. And yet, there are broader and even more interesting generalizations, those that transcend specific cognitive domains. Cognitive linguists have been especially attentive to this dimension of the new research, and they have argued persuasively for the cognitive generality of the mappings, correspondences, bindings, integration, perspectival organization, windows of attention, pragmatic functions, framing, force dynamics, prototype structures, and dynamic simulations that underlie the construction of meaning as reflected by language use. As a result, linguistics is no longer a self- contained account of the internal properties of languages; it is in its own right a powerful means of revealing and explaining general aspects of human cognition.

References

Bateson, G. 1972. Steps to an Ecology of the Mind. New York: Ballantine Books.

Coulson, Seana. 1997. Semantic Leaps: Frame-Shifting and Conceptual Blending. UCSD Ph. D. dissertation.

Cutrer, Michelle. 1994. Time and Tense in Narratives and Everyday Language. Doctoral dissertation, University of Calfornia at San Diego.

Dinsmore, J. 1991. Partitioned Representations. Dordrecht: Kluwer.

Fauconnier, G. 1994. Mental Spaces. New York: Cambridge University Press. [Originally published (1985) Cambridge: MIT Press.]

Fauconnier, G. 1997. Mappings in Thought and Language. Cambridge: Cambridge University Press.

Fauconnier, G. & E. Sweetser. 1996. Spaces, Worlds, and Grammar. Chicago: University of Chicago Press.

Fauconnier, G. & Turner, M. 1996. Blending as a central process of grammar. In Conceptual Structure, Discourse, and Language, Ed. Adele Goldberg. Stanford: Center for the Study of Language and Information [distributed by Cambridge University Press].

Fauconnier, G. & Turner, M. 1998. Conceptual Integration Networks. Cognitive Science.

Fauconnier, G. & Turner, M. (in preparation). Making Sense.

Fillmore, C. 1985. Frames and the Semantics of Understanding. Quaderni di Semantica VI.2. 222-253.

Freeman, Margaret. 1997. Grounded spaces: Deictic -self anaphors in the poetry of Emily Dickinson. Language and Literature 6:1, 7-28.

Goffman, E. 1974. Frame Analysis. New York: Harper and Row.

Goldberg, A. 1994. Constructions. Chicago: University of Chicago Press.

Hofstadter, D. 1995a. Fluid Concepts and Creative Analogies. New York: Basic Books.

Koestler, Arthur. 1964. The Act of Creation . NY: Macmillan.

Kunda, Z., D. T. Miller, and T. Clare. Combining social concepts: the role of causal reasoning. Cognitive Science 14. 551-577.

Lakoff, George. 1987. Women, Fire, and Dangerous Things (Chicago: University of Chicago Press), case study 1, pages 380-415.

Lakoff, George and Mark Johnson. 1999. Philosophy in the Flesh. New York: Basic Books.

Lakoff, George and Mark Turner. 1989. More than Cool Reason: A Field Guide to Poetic Metaphor (Chicago: University of Chicago Press).

Langacker, R. 1987/1991 . Foundations of Cognitive Grammar. Vol.I. ,II, Stanford University Press.

Liddell, Scott K. 1996. Spatial representations in discourse: Comparing spoken and signed language. Lingua 98:145-167.

Mandelblit, Nili 1997. Grammatical Blending: Creative and Schematic Aspects in Sentence Processing and Translation. Ph.D. dissertation, UC San Diego.

Moser, D. and D. Hofstadter. (undated ms.) Errors: A Royal Road to the Mind. Center for Research on Concepts and Cognition. Indiana University.

Nunberg, G. 1978. The Pragmatics of Reference. Bloomington, Ind.: Indiana University Linguistics Club.

Oakley, Todd. 1995. Presence: the conceptual basis of rhetorical effect. Ph. D. dissertation, University of Maryland.

Robert, Adrian. 1998. Blending in the interpretation of mathematical proofs. In: Discourse and Cognition: Bridging the Gap. Edited by Jean-Pierre Koenig. Stanford: Center for the Study of Language and Information (CSLI) [distributed by Cambridge University Press].

Sanders, J. and G. Redeker. 1996. Perspective and the Representation of Speech and Thought in Narrative Discourse. In Fauconnier, G. & E. Sweetser, eds. Spaces, Worlds, and Grammar. Chicago: University of Chicago Press.

Sweetser, E. 1990. From etymology to pragmatics: the mind-as-body metaphor in semantic structure and semantic change. Cambridge: Cambridge University Press.

Sweetser, Eve. Mental Spaces and Cognitive Linguistics: A Cognitively Realistic Approach to Compositionality. this volume.

Talmy, L. 1977. Rubber-sheet Cognition in Language. Proceedings of the 13th Regional Meeting of the Chicago Linguistic Society.

Talmy, L. 1978. Figure and Ground in Complex Sentences. In: Universals of Human Language: Syntax .(vol. 4). Edited by Joseph Greenberg. Stanford University Press.

Talmy, L. 1985. Force Dynamics in Language and Thought. Papers from the Parasession on Causatives and Agentivity. Chicago: Chicago Linguistic Society.

Talmy, L. 1998. Fictive Motion in Language and “ception.” In Paul Bloom, Mary Peterson, Lynn Nadel, and Merrill Garrett, eds., Language and Space. Cambridge, Mass.: MIT Press.

Turner, Mark. 1991. Reading Minds: The Study of English in the Age of Cognitive Science. Princeton: Princeton University Press.

Turner, Mark. 1996. The Literary Mind. New York: Oxford University Press.

Van Hoek, Karen. 1997. Anaphora and Conceptual Structure. Chicago: The University of Chicago Press.

Zbikowski, Lawrence. 1997. Conceptual blending and song. Manuscript, University of Chicago.

法语发音 http://phonetique.free.fr/indexphoncons.htm

可测试自己是否能分辨出法语音标间的差别, 是有声的, 很好玩!!

Accord (méthode de fran?ais) http://www.didieraccord.com/

初级练习 http://www.didieraccord.com/Accord1/dossier1/

“Accord” 是法语学习的教材, 进入它的网站有自我学习及订正的练习, 对於写作是一个练习的好机会.

主题学习 http://babelnet.sbg.ac.at/canalreve/bravo/index2.htm

站内依主题分, 共八个单元, 每个单元内又分会话, 字汇及写作. 会话的部份有小影片辅助教学, 生动活泼. 此网站比较适合已有法语基础者.

进阶练习 http://www.didierbravo.com/html/jeux/index.htm

法语程度不错者, 可进入挑战.

BBC http://www.bbc.co.uk/languages/french/index.shtml

实用法语课程互动式学习

用画学习法语还可以欣赏名画 http://peinturefle.free.fr/

不仅可以学到法文, 还可以欣赏到一些名画. 很棒的一个网站.

Polar FLE http://www.polarfle.com/

法语多媒体教学网站 http://www.laits.utexas.edu/fi/index.html

法英对照, 有许多实用字汇及影音, 初学者可多加利用.

法语动词变化 http://www.hku.hk/french/starters/lexis/fiche06_verb1.htm

法语动词变化很难但也很重要, 希望这个网站对学员们能有所帮助.

想取个法文名字吗? 可进入此网站查查 http://www.prenom.com/

http://www.lexode.com/keskidi/artiste52330titre.html

法国儿歌欣赏 :有歌词及歌唱语音播放. 相信学法语的您很快就可以朗朗上口.

小王子法文版网站 : http://galeb.etf.bg.ac.yu/mp/mp/pp.html

法国国铁 SNCF 网路订票 : http://www.voyagessncf.com/dynamic/_SvHomePage

法国航空公司网络订票: http://www.airfrance.fr/

推荐一个小语种学习网,其中有法语,很好的哦 http://www.languageguide.org/

http://www.languageguide.org/ 里面的法语单词都分门别类的还图文并茂法语单词读音也很好听也~~~ 可以学到很多基本的法语单词!

CCTV的法语频道可以在网上直接看原来CCTV的法语频道可以在网上直接看的，我到现在才发现。内容还挺多的。大家仔细找找还有法语音乐什么的，也可以在线看。

http://fr.cctv.com/francais/　这个是首页。

http://fr.cctv.com/program/journal/02/07/index.shtml

这个是可以在线观看的，速度还不错。顺便在讲一下，不知道是我浏览器的原因，还是什么原因，浏览CCTV的东西要刷新两三次，才能打开网站…… 全是免费的资源，音像，影像，文本教程一应俱全，有法语，西班牙语，德语还有意大利语。英语解说。

法语： http://www.bbc.co.uk/languages/french/

西班牙语： http://www.bbc.co.uk/languages/spanish/

学法语的网站,很不错的,英法对照噢 http://bbc.co.uk/languages

很多法语视频 还有很多音乐，电影片段 http://www.tv5.org/TV5Site/programmes/accueil_continent.php

http://www.bbc.co.uk/languages/french/

最好的该算是 BBC 的 French Steps 当然，这需要你有一定的英语基础，毕竟这是人家英国人学法语的网站，用flash做的，全对话场景教学，并有配套

教材供下载中国国际广播电台

法语频道 http://fr.chinabroadcast.cn/

中国网 法语频道 http://www.china.org.cn/french/index.htm

French Tutorial http://www.frenchtutorial.com/

http://www.rfi.fr

法国国际广播电台的官方网站，左侧 la langue francaise很好的，有文稿，是每日新闻，声音文件可以下载!一些法文法语学习网站