Gentner, Dedre (eds); Susan Goldin-Meadow;
Language in mind: advances in the study of language and thought
MIT Press, 2003, 528 pages [gbook]
ISBN 0262571633, 9780262571630
topics: | language-acquisition | whorfian | usage
Considers the Whorfian hypothesis from many perspectives. Sapir-Whorf hypothesis: languages vary in their semantic partitioning of the world, and the structure of one's language influences how one understands the world. Thus speakers of different languages perceive the world differently. In the introduction, Gentner and Goldwin-Meadow outline the controversies surrounding this view: For the last two decades, the hypothesis that language can influence thought—generally known as the Whorfian hypothesis — has been in serious disrepute. Admitting any sympathy for, or even curiosity about, this possibility was tantamount to declaring oneself to be either a simpleton or a lunatic. The view of most language researchers is well expressed by Pinker (Language Instinct 1994, 65): Most of the experiments have tested banal ‘weak’ versions of the Whorfian hypothesis, namely that words can have some effect on memory or categorization. Some of these experiments have actually worked, but that is hardly surprising. Devitt and Sterelny (1987, 178) express this skepticism even more strongly: [T]he argument for an important linguistic relativity evaporates under scrutiny. The only respect in which language clearly and obviously does influence thought turns out to be rather banal: language provides us with most of our concepts. They then suggest that arguments like the latter "exemplifies a rather schizophrenic way" of viewing Whofrianism: on the one hand, the language-influences-thought is considered unimportant, yet in the same breath it is stated that language provides us with most of our concepts. the latter is actually stronger than even the most pro-Whorf researchers.
Whorf was not the first to express the idea that language influences thought. For example, Humboldt (1836) viewed language as the formative organ of thought and held that thought and language are inseparable (see Gumperz and Levinson 1996a; Lucy 1996, for reviews). Whorf’s own views were somewhat more subtle than is generally realized. Along with his well-known strong conjecture: We dissect nature along lines laid down by our native language. The categories and types that we isolate from the world of phenomena we do not find there because they stare every observer in the face; on the contrary, the world is presented in a kaleidoscope flux of impressions which has to be organized by our minds—and this means largely by the linguistic systems of our minds. (1956, 213) he also considered weaker views: My own studies suggest, to me, that language, for all its kingly role, is in some sense a superficial embroidery upon deeper processes of consciousness, which are necessary before any communication, signaling, or symbolism whatsoever can occur ... (1956, 239) Nonetheless, the hypothesis that has come to be known as the Whorfian hypothesis, or alternatively the Sapir-Whorf hypothesis, states that (1) languages vary in their semantic partitioning of the world; (2) the structure of one’s language influences the manner in which one perceives and understands the world; (3) therefore, speakers of different languages will perceive the world differently.
Why would anyone ever come up with the hypothesis that the language we speak shapes the thoughts we think? Consider a plausible scenario. When retelling an event, speakers of Turkish are required by their language to indicate whether they themselves actually witnessed that event (Aksu-Koc¸ and Slobin 1986). Of course, the speaker knows whether she witnessed the event. However, she may not be interested in conveying this bit of information to the listener. Speakers of English have the option (which they often exercise) of leaving out whether they actually witnessed the event they are retelling—speakers of Turkish do not. After many years of routinely marking whether they witnessed an event, it is possible that Turkish speakers will tend to encode whether an event has been witnessed, whether or not they are talking. That is, Turkish speakers may habitually attend to this feature of the world much more than English speakers do. In other words, their way of viewing the world may have been altered just by becoming speakers of Turkish as opposed to English. This is the kind of reasoning that underlies the Whorfian hypothesis.
Language evokes ideas: it does not represent them. Linguistic expression is thus not a straightforward map of consciousness or thought. It is a highly selective and conventionally schematic map. At the heart of language is the tacit assumption that most of the message can be left unsaid, because of mutual understanding (and probably mutual impatience). (Slobin 1979, 6) Add to this the fact that what is conventionally schematic in one language may not be so in another. Effectively, Slobin here echoes Whorf: Users of markedly different grammars are pointed by their grammars towards different types of observations and different evaluations of externally similar acts of observation.... ([1940] 1956, 221) Whorf in turn follows Boas, who pointed out that those elements in a language that are obligatory—the grammatical categories—are what determine ‘‘those aspects of language that must be expressed’’ (1938, 127). In short, what is obligatory in each language can differ, so that speakers only express part of whatever they have in mind (Boas 1911). What this implies is that speakers will select different details, different aspects, from their representations of each scene or event, depending on what language they are speaking.
Language and Mind: Let's get the issues straight [p. 25-46] Simple Nativism: - Tenet1: syntax of lg is fundamentally universal and innate [Chomsky] - Tenet2: semantics is given by an innate "language of thought" - a position ably defended by Fodor:1975 - The central property is that all properties of lg are dictated by inbuilt mental apparatus. - Not associated with any adaptationist or evolutionary argument for lg. biological basis of semantics - addresses an endowment in culture, which is outside biology. the conceptual packages that are coded in any lg are a product of culture, and NOT of biology. e.g. infant speech perception : by six months, this ability has been warped into line with the local lg they are hearing. [monkeys and other mammals can't do this]
Nativist view: infants are built to expect lg diversity, and have special mechanisms for "tuning in" [Kuhl Metzoff 96,97]. Protagonists include linguists like Jackendoff (see Landau/J:1993), cog psyc (Pinker:1994, Gleitman - see Li/Gl:2002) and the so-called evolutionary psychologists like Tooby and Cosmides 1992. p. 27: Choi and Bowerman:00; McDonoughetal,thisVol.: at age 9 mo, infants make both Korean and English spatial distinctions; by 18 mo: tuned into the local lg McDonough, L., Choi, S., and Mandler, J. (in press). Understandhg spatial relations: Flexible infants, lexical adults. Cognitive Psychology.
p.28: Simple Nativists hold that linguistic categories are a direct projection of universal concepts that are native to the species: Knowing a language, then, is knowing how to translate mentalese into strings of words and vice versa. People without a language would still have mentalese, and babies and many nonhuman animals presumably have simpler dialects. (Pinker 1994, Language instinct. Morrow. p.82) Learning a language is on this view simply a matter of learning the local projection, that is, finding the local phonetic clothing for the preexisting concepts. Or as Li and Gleitman (2002, 266) put it: - Language has means for making reference to the objects, relations, properties, and events that populate our everyday world. It is possible to suppose that these linguistic categories and structures are more or less straightforward mappings from a preexisting conceptual space, programmed into our biological nature: Humans invent words that label their concepts. Hence, they hold, "the grammars and lexicons of all languages are broadly similar."
In these situations, some kind of angular specification on the horizontal plane is called for - as in 'The ball is behind the tree'. It turns out that although languages vary greatly in the detailed geometry employed, there are three main families of solutions: an egocentric (or more accurately viewpoint-dependent) relative system (as in the 'The ball is left of the tree'), a geocentric absolute system (as in 'The ball is north of the tree'), and an object-centered intrinsic system (as in 'The ball is at the front of the truck'). These three are all polar coordinate systems and constitute the best claim for universals in the spatial domain. But there are some important caveats. First, not all languages use all three systems. Rather, they form an inventory from which languages must choose at least one-all combinations are possible, except that a relative system entails an intrinsic system. That means there are languages without words for 'left' or 'right' directions, but where all spatial directions must be specified in terms of cardinal directions like 'east' (so one has to say things like 'Pass the northern cup', 'There's a fly on your northern leg', etc.).
To sum up: the Simple Nativist idea (as voiced by Pinker and Gleitman) that universal concepts are directly mapped onto natural language words and morphemes, so that all a child-learner has to do is find the local name as it were, is simply false. There is an ideological overtone to Simple Nativism: the independence of thought from language opens up to us the freedom of will and action ("[Slince mental life goes on independently of particular languages, concepts of freedom and equality will be thinkable even if they are nameless" Pinker 1994, 82). ... no one, not even Whorf, ever held that our thought was in the infernal grip of our language. Whorf's own idea was that certain grammatical patterns, through making obligatory semantic distinctions, might induce corresponding categories in habitual or nonreflective thought in just the relevant domains (see Lucy 1992b for careful exposition). ... There are languages with no or very few number words, and without a generative system of numerals - it seems unlikely that the speakers of such a language would ever entertain the notion 'seventy-three', let alone that of a logarithm, and certainly their fellows would never know if they did. [IDEA: place-value system --> generative numerals] 2. Fodor himself adopts the only way out of this dilemma, which is to say that every lexical concept in every language that ever has been and ever will be is already sitting there in our heads. So Cro-Magnon man already had the notions 'neutrino' and 'piano', but probably hadn't gotten around to giving them phonetic form!
Edward Munnich and Barbara Landau, The effects of spatial lanuage on spatial representation: Setting some boundaries, p. 113-155
Does learning one's language affect the foundational nonlinguistic spatial representation upon which the language of space stands? English "on" vs. A. German "an" : support involving attachment, - "painting on the wall", or "tab on the soda can" B. German "auf": support without attachment, - "cup on the table" Japanese ue (-ni) [literally, top (locative)], refers to both "ON" and "ABOVE" Bengali: upor - also both "on" and "above" Hindi: upar - can be both, but "par" is more usual for "on" Korean kkita ~ English "put on" - tight-fit, "put cap on pen" ppusta - "loose fit" - "put blanket on bed" or "book on table" Fig. on p.119, from Bowerman 1996, excellent figure of three cases: A: apple-on-table, B: door-on-cupboard, C: apple-in-bowl A B C English: { ON } IN Finnish: -LLA { -SSA } Dutch: OP AAN IN Spanish { EN } Hindi { -par } -mein Bengali upar -te -te
Werker and Tees 1984: 6-8 mo. children from native English bkgd, can distinguish dental t from retroflex T, a distinction unimp in English, using a conditioned head-turn paradigm. claim: This ability to distinguish is lost by adulthood Werker and Logan 1985: if contrastive sounds that are not a phonemic distinction in English are presented, Eng adult speakers can discriminate as well as native Engl phonemes, if gap (interstimulus interval) is < 500 ms. If ISI > 1500 ms, can't. Hence, there may be two diff modes of processing same stimulus, "acoustic" vs. "phonemic" mode. thus linguistic experience reorganizes phonemic distinctions, but does not lead to permanent loss of the ability to distinguish non-native acoustic contrasts. 124
Berlin and Kay 1969: "basic" color terms in lgs, varies from 2 (Dani) to 12; English has 11. Heider and Olivier 1972 tested if num of colour terms affects recognition. Showed colour chips varying in hue and brightness, but all with low saturation, to Dani and English speakers. After a first chip was shown, after 30 second unfilled interval, they were to pick a chip from a set of similar colors. Hypothesis was that "verbal code would interact with the visual to influence the nature of memory errors" (p.339). But though the lg names were roughly 11 and 2, the colour matches were roughly the same, both groups showed nearly identical patterns of confusion in memory. even "an image tagged "mola" by Dani S(ubject) might ahve shifted in any direction so long as it remained withint the color space recognized as mola by that S. By any of these mechanisms, colors which were closer to each other in the name structure should have been more often confused with each other in recognition... We did not find such an effect" p. 351-2 Heider, E., and Olivier, D. 1972 The structure of the color space in naming and memory for two languages. Cognitive Psychology, v.3, 337-354 6.4.1 Recently H&O's findings have been disputed in a study by Robertson, Davies and Davidoff 2000, but we disagree with their interpretation and believe that they are not nec incompatible with H&O's conclusions. In RDD, they use the lg Berinmo instead of Dani. Dani has two classes, orange, yellow, red (roughly "warm"), and blue, green (roughly, cool). Thus the Dani categories are supersets of Engl categories. However Berinmo cuts across - wor covers some green, as well as some yellow/orange. Using the same methods as H&O, they report that the fit between Berinmo names and Berinmo color memory is better than the fit between Berinmo color memory and English color memory; hence there is a effect of lg. But one would expect that lifelong use of colour names would result in bias "within a single culture" - and not across cultures. In any event, both Dani and Berinmo exhibit poorer memory compared to English.
debate: cross-linguistic diff's in spatial lg --> does it alter cognition? As the visual image fades after a very short time, the encoding of location may become biased or even converted into verbal code, which would change the contensts of memory. Hayward and Tarr: 1995: showed circle around a computer - and had them a) NAMING: fill in the blanks in : The circle is ___ the computer. People filled in above, below, left, right for close to principal axes, and usage of these terms declined as the position was further from these. b) MEMORY: after showing stimulus, they saw a visual mask. Then a test array appeared in which the position of the circle was either the same or had a small displacement. They were to say if it was same or not. Here they found that participants' memory was best when it was in one of the principal axes. Suggested that the principal axes organized both memory and also lg. Munnich, Landau and Dosher:2001: replicated with circle around square; tested on both Japanese and English (both have terms for princ axes; japanese does not distinguish contact, between on and above). munnich-landau-doshercogn01spatial-language-reprresentation-cross-linguistic.pdf Found: both memory and naming were more accurate near principal axes. But not much diff between contact / non-contact. Thus no effect of lg here.
see Malt Sloman, Gennari:2003 (this volume): examined containers - Chinese, Spanish, and English speakers named 60 types of containers - one gallon milk jug, peanut butter jar, baby bottle, etc. --> surprising variation e.g. Spanish frasco (or its diminutive frasquito) covered 28/60 objects, which in Engl were named bottle (6), container (3) and jar (19). Chinese most common name-word (40) spanned English bottle (13) container (8) and jar (18). Next did sorting tasks. Asked to sort on 3 criteria: - physical sort based on physical qualities - appearance, material, etc. - functional sort based on how it contains the substance - in a stack, sep pieces, single solid, etc. - overall sort based on overall qualities of continer, what it looks like, what it[s made of, how it contains any substance, etc. Naming similarity (roughly - degree to which any pair of objects have similar name distributions) was .35 to .55. Sorting similarity correlations: .91 to .94 for overall similarity; .82 to .89 for phys similarity, and .55 to .79 for function. Thus all are substantially higher, and possibly engaged a diff kind of similarity metric than in naming task...
Berman/Slobin:1994, Talmy:1985,1991 Talmy: boat floated into the cave: float into can be separated: Spanish: floated and entered most commonly conflated: motion-manner, motion-path, and motion-figure English: motion-plus-manner: roll, run, float, slide motion-plus-path: enter, exit, climb motion-plus-figure: rain, spit English: tend to conflate motion+manner, keep path separate Spanish: tend to conflate motion+path, keep manner separate
Gennari etal:2000 Three groups shown a video with motion in which manner+path varied. e.g. man CARRYING a board INTO a room A. asked to describe as shown B. not asked; could verbalize if they wanted C. given a shadowing task, so could not verbalize. Then were asked in 1. RECOGNITION task between slight variants in manner or in path, not both. e.g man CARRYING board OUT of the room, or DRAGGING board INTO the room. 2. SIMILARITY task: two events shown simult, more similar one is to be chosen, lures varied in manner or path, not both. 3. DESCRIPTION task: verbally describe Whether people mentioned manner or path in the descriptions depended on lg. group A, Spanish speakers who tend to conflate motion and path, tended to see similarity more in terms of path than English speakers. Thus, effect of language is seen mainly where language is already used to describe the scene, linguistic determinants do matter.
I. Introduction 1. Whither Whorf Dedre Gentner and Susan Goldin-Meadow II. Position Statements 2. Languages and Representations Eve V. Clark 3. Language and Mind: Let's Get the Issues Straight Steve Levinson 4. The Key is Social Cognition Michael Tomasello III. Language as Lens: Does the Language We Acquire Influence How We See the World? 5. Sex, Syntax, and Semantics Lera Boroditsky, Lauren A. Schmidt, and Webb Phillips 6. Speaking vs. Thinking About Objects and Actions Barabara C. Malt, Steven A. Sloman, and Silvia P. Gennari 7. The Effects of Spatial Lanuage on Spatial Representation: Setting Some Boundaries Edward Munnich and Barbara Landau 8. Language and Thought Online: Cognitive Consequences of Linguistic Relativity Dan I. Slobin If I tell you about my "friend" in English, you will expect that sooner or later you will discover the sex of the friend, because you know that third-person pronouns in English indicate gender. If I go on and on to refer only to "my friend" or "they" you will begin to suspect that I have reason to conceal the person’s gender. However, if we have the same conversation in a language that has no gendered pronouns, such as Turkish or Chinese or Hungarian, you probably will not have such suspicions. IV. Language as Tool Kit: Does the Language We Acquire Augment Our Capacity for Higher Order Representation and Reasoning? 9. Why We're So Smart Dedre Gentner 10. Does Language Help Animals Think? Stan Kuczaj and Jennifer L. Hendry 11. What Makes Us Smart? Core Knowledge and Natural Language Elizabeth Spelke 12. Conceptual and Linguistic Factors in Inductive Projection: How Do Young Children Recognize Commonalities Between Animals and Plants? Kayoko Inagaki and Giyoo Hatano 13. Language For Thought: Coming to Understand False Beliefs Jill G. de Villiers and Peter A. de Villiers V. Language as Category Maker: Does the Language We Acquire Influence Where We Make Our Category Distinctions? 14. Space Under Construction: Language-Specific Spatial Categorization in First Language Acquisition Melissa Bowerman and Soonja Choi 15. Reevaluating Linguistic Relativity: Language-Specific Categories and the Role of Universal Ontological Knowledge in the Construal of Individuation Mutsumi Imai and Reiko Mazuka 16. Interaction of Language Type and Referent Type in the Development of Nonverbal Classification Preferences John A. Lucy and Suzanne Gaskins 17. Thought Before Language: Do We Think Ergative? Susan Goldin-Meadow
The idea that the language we speak influences the way we think has evoked perennial fascination and intense controversy. According to the strong version of this hypothesis, called the Sapir-Whorf hypothesis after the American linguists who propounded it, languages vary in their semantic partitioning of the world, and the structure of one's language influences how one understands the world. Thus speakers of different languages perceive the world differently. Although the last two decades have been marked by extreme skepticism concerning the possible effects of language on thought, recent theoretical and methodological advances in cognitive science have given the question new life. Research in linguistics and linguistic anthropology has revealed striking differences in cross-linguistic semantic patterns, and cognitive psychology has developed subtle techniques for studying how people represent and remember experience. It is now possible to test predictions about how a given language influences the thinking of its speakers. Language in Mind includes contributions from both skeptics and believers and from a range of fields. It contains work in cognitive psychology, cognitive development, linguistics, anthropology, and animal cognition. The topics discussed include space, number, motion, gender, theory of mind, thematic roles, and the ontological distinction between objects and substances.