An action suit, not a straightjacket: Whorf on language, Guy Deutscher on Whorf

Guy Deutscher, Through the language glass – Why the world looks different in other languages (2010) [1] – reviewed by Nick Enfield.

In April of 1985, Texas Tech student and rape victim Michele Mallin made about as consequential an error of cognition as one can imagine, when she mistakenly identified 26-year old Timothy Cole as her attacker, first from a photograph, then in a police line-up. Cole would later die in prison after serving 13 years of a 25-year sentence for a crime he didn’t commit. Many factors are known to contribute to such unthinkable distortions of memory, but one may seem surprising: the role of language.

The science of memory has shown that if we verbally describe something we have seen, as when a victim first describes an attacker to police, this verbal representation can ‘overshadow’ our exact recollection of the original experience. By talking about something, we can become more likely to make a wrong decision based on what we said (and now think) we saw. When psychologists Jonathan Schooler and Tonya Engstler-Schooler discovered this effect for faces, they also found that verbal description can make memory worse for another perceptual domain: colour. Their test subjects were shown a set of subtly different shades, and were worse at later identifying the one they had seen if they had verbally described it. Those who had not verbally described the original colour, on the other hand, were better at remembering it. The authors concluded: ‘Some things are better left unsaid.’

How is it that the verbal description of an experience can cause a distortion of memory?

This detrimental effect of language on cognition is an unwanted by-product of an otherwise critical function of language: categorization. Categorizing things means grouping them in a single class, treating them as effectively identical by disregarding—or discarding—incidental differences. Language is centrally involved. The concepts that words encode are ready-at-hand devices for categorizing. So, for example, while no two chairs are the same, we are treating them as the same when we refer to them both with a single word chair. Of course the man who says ‘I bought two new chairs today’ knows whether they were different colours, but his addressee can’t know it just from hearing the phrase ‘two chairs’. The linguistic category omits distinguishing information, and the distinctions made or omitted will be different in different languages. English won’t let you include a stool in the ‘chair’ class, though many languages will. In Lao, a language spoken in Laos, the word tang covers chairs and stools in a single idea. A comparison of any two languages will quickly yield hundreds of examples like this.

The trusted concepts that words provide give our minds ways to economize on unnecessary detail. The trade-off is that while we may save on mental processing, for example by not having to remember a chair’s peculiarities once we have classified it as a chair, we may lose later access to context-specific details of mental representation. On balance, the trade-off is good. But in practice the finer details of an image are not always dispensable, as we learn if we find ourselves buying the wrong tint of paint for the house, or, more consequentially, convicting an innocent man.

If language can imperceptibly constrain or channel our thoughts, this raises a deeply counter-intuitive challenge to our sense of free will. Do our trains of thought run on pre-laid tracks? And could those tracks be linguistically laid, at least in part, such that with each different language our thoughts are transported to different destinations? A lineage of language scientists has posed and explored these questions, from Herder and Humboldt to Boas and Sapir and beyond. Perhaps the best known is American linguist and anthropologist Benjamin Lee Whorf.

Whorf’s considerable accomplishments in Native American and Mesoamerican linguistics, including his 1930s fieldwork on the Aztec language and his contributions to the decipherment of Mayan hieroglyphs, are still referenced in specialist literature of that area. In linguistics more generally, despite some gaffes such as his flawed description of time expressions in the Hopi language, Whorf made lasting contributions to the study of how meaning is encoded, both overtly and covertly, in grammatical systems. But Whorf is most discussed for his proposals concerning relations between language, thought, and culture. In a short article written in 1939, titled ‘The relation of habitual thought and behaviour to language’, Whorf considered the possible role of language in human error leading to accidental fires, drawing on his professional experience as an insurance investigator. What he pointed to was not unlike the verbal overshadowing effect for faces and colours. In line with previous proposals by early Twentieth Century German-American anthropologists Franz Boas and Edward Sapir, Whorf argued that a linguistically grounded habit of thinking might play a causal role in non-linguistic behaviour. In one of his fire insurance examples, an explosion occurred when workers were careless with cigarettes near empty gasoline drums. He reasoned that if an English speaker thinks of a gasoline drum as ’empty’ because there’s no gasoline in it, this use of the word ’empty’ suggests to the thinker that there’s nothing else in the drum either. Hence it wouldn’t enter one’s mind that the drum is in fact full of vaporous fumes, and hence the dangerous behaviour.

If you think this means Whorf was taking language to be the sole determinant of behaviour, or even a straitjacket for thought, you would not be alone. But Whorf never said these things, nor did he think them. Presenting his case in the ‘condensed and unqualified form’ required for that short article, he had hoped the reader ‘would use his thinking apparatus’ and adjust accordingly. [2] In response to his editor’s suggestion that he may have over-emphasized the role of language in channelling behaviour, Whorf explained:

I have thought of possibly adding a brief statement or a footnote saying that I don’t wish to imply that language is the sole or even the leading factor in the types of behaviour mentioned such as the fire-causing carelessness through misunderstandings induced by language, but that this is simply a coordinate factor along with others. It didn’t seem at first that this should be necessary if the reader uses ordinary common sense, but then one never can tell.

Since Whorf gave his readers the benefit of the doubt, let us extend the same courtesy. What he suggested was that when deciding how to behave, one might naturally use language in thinking, and through this one may fail to recognize important information that happens not to feature in a strictly linguistic rendering of the state of affairs at hand. It’s not that people can’t comprehend alternative renderings of reality. Nor is it that they can’t think without language. It’s that we are creatures of cognitive and linguistic habit. There are things we habitually pay attention to, things we habitually overlook, and ways we habitually reason in deciding how to act.

A common gloss of Whorf’s view is that a given language makes us more likely to think certain things, but one could also put it that a language makes us more likely not to think certain things. The idea that we habitually disattend to much of what goes on around us is now well established and fairly well understood. Research on rational heuristics in decision-making shows that to routinely disregard a portion of the information available is not only a common strategy, it also makes good sense. The logic of decision-making includes an effort to minimize the costs involved. It’s the same from small things like choosing among brands of breakfast cereal to big things like selecting a marriage partner. Once you’ve locked on to the problem, your next step is to find ways to narrow the search for an appropriate solution and ‘lock off’ by making a decision whose outcome shows the best balance of optimal benefit for lowest cost. And a decision should be made without unnecessary delay, so one can get on with life—that is, with more decisions.

Take a simple decision like what to order at a restaurant. One strategy would be to study the entire menu and weigh up all the options, comparing them on some criteria and selecting the best. But such thoroughness could be cognitively costly, not to mention distracting from dinner conversation. A good alternative is the strategy known as satisficing: Don’t waste your time studying all the options, simply lock off at the first solution that is good enough for current purposes. How does one determine what qualifies as good enough? One turns to concepts as devices for behavioural categorization. Concepts are important in decision-making because they give us criteria for recognizing instances of what we are looking for. Whorf’s point was that language is likely to play a central role because the most ready-at-hand concepts are the ones encoded in the languages we speak.

All of this suggests a prima facie argument for a form of cultural relativity grounded in differences between languages. Given that concepts provide a basis for categorization and decision-making, and given that different languages supply their speakers with different concepts, then different languages provide their speakers with different bases for decision-making, and, subsequently, different patterns of behaviour. It’s a captivating possibility. We are members of a single species, but could it be that the different linguistic systems we inherit from different cultural histories cause us to think and act in fundamentally different ways? Or that as bilinguals, when we switch languages, we switch cognitive personalities? This is what linguistic relativity suggests.

Some may recoil at the idea of relativity in any form, appealing to plain common sense. Philosopher Paul Boghossian opens his 2006 anti-relativist book Fear of Knowledge with a controversy surrounding Native American origins. While many archaeologists believe there was a migration ten millennia ago from Asia into America via the Bering Strait, some Native Americans believe that their people emerged onto the Earth from a subterranean spirit world. Are these two versions of reality equally valid, as some relativists claim? Absolutely not, says Boghossian. There are facts which do not stop being facts simply because we believe something else. Rocks will sink, whether you like it or not. True enough, but a politically correct doctrine of equal validity is not the only version of relativity.

A version of relativity more likely to succeed begins with the observation that we cannot determine facts independently from the measuring instruments that are used. In life, our measuring instruments are our bodies. Why can dogs hear sounds we can’t? Because dogs have different bodies from us. For humans with human ears, those ultrasonic noises may as well not be real (though we are able to infer their existence using other means, from hi tech instruments like spectrograms to low tech measures like naked-eye observations of dogs’ behaviour). The body defines an individual’s horizons, both limiting and licensing our possible perceptions and actions. If you have the body of a bat, a pitch-dark cave will seem like a good place to be. But with the body of an earthworm, you will feel at home in a stretch of turf. If these are not different worlds they are certainly different worldviews.

Along similar lines, you can think of each language as a kind of body for thinking and acting, and this is not meant metaphorically. Thinking is a dynamic, active thing, and seldom wholly internal. Philosopher of mind Gilbert Ryle advocated a focus on the public side of cognition, defining it not by putative internal states but by the ‘assemblage of performances’ that such states enable. To perceive or understand the world is to relate to it, interpret it, and react to it. The reasoning involved draws on categories in a range of ways, and many of our categories are supplied by the languages we speak. The mind, including its learnt linguistic components, is a purpose-built tool kit for cognitive action, just like a body is a tool kit for physical action. And because languages are so differently structured, each one is like the body plan of a different species, affording its users different ranges of possibility. A language, then, doesn’t imprison your mind, it equips it. It’s not a straitjacket, it’s an action suit.

This is pretty much what Whorf and his predecessors were suggesting. In the words of Edward Sapir, Whorf’s teacher at Yale, ‘the language habits of our community predispose certain choices of interpretation.’ [3] On this view, the study of different languages, especially of those least like so-called Standard Average European, has the potential to supply us with new and different ways for actively interpreting the world. Whorf avidly promoted the study of lesser-known languages, such as those of Native America, since each provides an opportunity to broaden our worldviews in one more way. While nobody is ‘free to describe nature with absolute impartiality’, Whorf argued, the person who would come closest ‘would be a linguist familiar with very many widely different linguistic systems.’ [4] Today’s language explorers are fulfilling this ambition: as linguist Nicholas Evans writes in his 2009 book Dying Words, ‘we study other languages because we cannot live enough lives’. [5]

*

How are we to treat Whorf’s idea today, over 70 years on? Some have taken it to be no more than a suggestive framework, serving as a conceptual guide for anthropological research or as a source of hypotheses for systematic psychological testing. Others have pressed too hard, dismissing the whole idea for its failure to take the weight. But it is not clear that Whorf’s work was designed to bear that load. Nor is the target of critique always the same as what Whorf actually proposed. Cognitive anthropologist Stephen Levinson describes the problem: ‘It is as if the topic of “Whorfianism” is a domain where anybody can let off steam, go on mental holiday, or pounce upon an ideological enemy’. [6] Whorf commentator Penny Lee counts the ways he has been ‘misread, unread, and superficially treated’, [7] not to mention ridiculed: epithets range from ‘amateur’ to ‘insurance adjuster’ to ‘fundamentalist Christian’.

Linguist Guy Deutscher is the latest to join the trend of Whorf-derision in Through the Language Glass. His scorn for Whorf and his work is expressed with words like ‘con man’, [8] ‘bogus’, [9] ‘bankruptcy’, [10] and ‘delusion’ [11]. This is puzzling, as Deutscher’s book promotes a decidedly Whorfian form of linguistic relativity, that is, where different cognitive effects are caused by different habits of speaking. At least one outcome of Deutscher’s prominent dismissal of Whorf may be the placation of those who would otherwise reject a relativist line of argument from the outset. After all, some of the more vehement critics of linguistic relativity have loud voices. Linguist Steven Pinker says it is ‘wrong, all wrong’, [12] while philosopher Jerry Fodor declares of relativism more generally, ‘I hate relativism more than I hate anything else, excepting, maybe, fibreglass powerboats’. [13] Very funny, but actually this is serious, so we must turn to matters of substance.

Established effects of language on thought such as the verbal overshadowing of memory are normally described as effects of language, not of languages. But consider such effects in light of the considerable structural and semantic diversity known to exist among languages of the world. Recall that the verbal overshadowing effect only occurs when associated with an experience ‘that defies complete linguistic description’. [14] What exactly are the kinds of experience that defy complete linguistic description? Crucially, the answer to this question is not absolute, but linguistically relative. It depends on the language you are speaking. As is increasingly apparent due to a growing body of evidence from intrepid linguistic fieldwork, the world’s thousands of languages vary significantly if not radically in terms of precisely what they enable their speakers to completely or readily describe.

Deutscher draws attention to this linguistic diversity when he critiques the old textbook cliché that all languages are equally complex. Actually, we can’t say with confidence whether they are equally complex or not, since, despite increasingly sophisticated attempts to address this question, there is no consensus view as to what linguistic complexity is, or how to measure it. (One then wonders how to support Deutscher’s assertion that languages aren’t equally complex.) But we can be sure that languages are complex in different ways. While English grammar has markings for tense and other temporal distinctions on the verb, Lao does not. Lao, for its part, has a highly differentiated grammatical system of pronouns, distinguishing between multiple levels of formality and politeness, while the English system is extremely simple. Despite the massive catalogue of such differences among languages, Deutscher nevertheless believes in a principle of equal expressiveness, an idea that literally ‘any thought can be expressed in any language’. [15] If this strong contention were true (though it is no better established than the claim of ‘equal complexity’ among languages), some might take it to be fatal for linguistic relativity, reasoning that if anything can be expressed in any language, how could two languages channel our thoughts differently in any significant sense?

A way forward is not to focus on what can be expressed in languages, but to focus instead on a known point of difference, namely on what languages must express, or what they most easily or habitually express. This approach was suggested a century ago by Boas, inspired by his field investigations of Native American languages with their very different grammatical demands. Many linguists since have been similarly intrigued by the sense in which different languages put their speakers under different obligations when speaking. The issue has also been explored by psychologists of language with an interest in the cognitive implications of online language processing. If my grammar requires me to explicitly mark plural versus singular, say, then I should monitor for ‘plurality’ in everyday experience more closely than a speaker of a language in which there is no such grammatical requirement. I should do this because later I may need to talk about the experience, and in that case I will need to fill in the critical blanks my grammar requires. This language-specific effect on general attention and memory was dubbed ‘thinking-for-speaking’ some 20 years ago by Berkeley psychologist Dan Slobin. [16]

The thinking-for-speaking effect is more widely acknowledged to be a genuine case of linguistic relativity than the kinds of general worldview effects proposed by Sapir and Whorf, but it is often regarded as a ‘weaker’ version, since it is thought by some to be restricted to specifically linguistic cognition. It is not at all clear, however, that these effects are purely linguistic. If one’s attention to a semantic domain is always heightened—because one never knows whether one might later have to put an experience into words—then this is surely a pervasive and general effect, beyond strictly linguistic cognition.

While Deutscher advocates a focus on what is grammatically obligatory or habitual, he does not mention Slobin’s work or develop similar implications. (John Lucy’s work is also inexplicably omitted, among many others.) What Deutscher does talk about is a trio of examples in which language variation has been shown to affect thought: colour terms, grammatical gender, and ways of referring to spatial location. Deutscher relates these findings from recent scientific literature: Our visual perception of colour distinctions can be affected by cross-language differences in colour vocabularies; Our implicit evaluations of inanimate objects like bridges can be affected by whether the relevant word is grammatically feminine (as for ‘bridge’ in German) or masculine (Spanish); Our reasoning about spatial relations can be affected by the particular frame of reference that is dominant in a given language (cardinal directions in some languages versus left/right in others).

Deutscher’s review is a lean treatment of some highlights from a complex scientific field. Many of the key studies and contributors to linguistic relativity research, past and present, go without mention. A further limitation in scope arises from Deutscher’s orientation toward the referential functions of language. While psychologists looking at relativity effects typically focus on language as a ‘means for making reference to the objects, relations, properties, and events that populate our everyday world’, [17] as Peggy Li and Lila Gleitman put it, there is much more to language than this referential function. Language is also, and no less importantly, a means for aesthetic and emotive expression, for making social contact, and for maintaining human relationships. These are not the same kinds of functions as making reference, but in these domains we should similarly expect different languages to provide different potentialities for different lines of interpretation and response. Through the ‘gossip and grooming’ functions that language has, we convince, cajole, and coerce others, we form alliances and cliques. Considering how deeply consequential these functions are for all involved, it is surprising how little we know about the differences between languages in the domains of social action, speech acts, conversational structures, modes of phatic communion. So, while Deutscher’s treatment of the relative effects of language on thought is a thin slice of the full range of research in the referential domain (beyond his three cases there is ample work on the linguistic expression of motion and locomotion, verbs of placement, logical relations, number systems, grammatical classification, mental predicates, object individuation, and more), all of that work is itself just part of a broader effort to understand range of ways in which languages may act as mirrors, lenses, action suits, or the like, explored within the linguistic sub-field of anthropology. Readers with an existing interest in the language-culture-cognition nexus may be hoping for a synthesis, or at least a wide-ranging review of linguistic relativity research. But there is still no one-stop shop.

Through the Language Glass is better viewed as a fun foray by a clever writer. The first half in particular features a wonderfully engaging historical account of scientific investigation of the diversity of colour perception in humans, and of colour semantics in different languages. The book would have been a winner had the author stuck to the readable narrative of the history of science that he does so well. Part of the pleasure is Deutscher’s positive compensation for modern science’s Anglocentrism and modernism by delving into 19th and 20th century German and French research reports that are hardly cited or discussed in more recent literature.

*

Linguistic diversity is causally related to cultural and cognitive diversity in a range of ways, and while some effects of language on thought are widely acknowledged, their possible interaction with the diversity of human languages is under-studied, yet likely to bear fruit. We don’t know if effects like verbal overshadowing are the same for all members of our language-using species, partly because we are still mapping the real extent of semantic and grammatical variation. Imagine that linguists discovered a language with a vocabulary for faces so perfect that you could recognize a man from the verbal description alone. An attacker’s face would no longer ‘defy complete linguistic description’, with the result that words would not overshadow memory for faces in that language, and the possibility of lower rates of wrongful conviction based on witness misidentification in that community. Linguistic fieldwork may yet find such a language, with a fine vocabulary for faces, but in the meantime we know that English is not it.

Not all of the known effects of language on thought are as consequential as the witness misidentifications that can destroy people’s lives. And not all linguistic effects are as seemingly innocuous as the subtle perturbation that colour vocabulary can have in one half of our visual field, or the subconscious sexual stereotyping of inanimate objects like keys or bridges in association with the grammatical gender of words for these things. Those who only admit a ‘weak’ version of linguistic relativity sometimes add that the weak version is trivial, or uninteresting. There is of course no independent measure of what we ought to find interesting. But trivial? An effect of language on thought is an effect of language on thought. If it means a speaker of one language is more likely to come home from the hardware store with the wrong shade of paint, then maybe that’s small potatoes, but not if it leads to a domestic argument. Nor if it makes a worker more likely to cause a factory fire, or a witness to put a man in prison for a crime he didn’t commit. If you think the effects of language on thought are trivial, try telling that to the ghost of Tim Cole.


Endnotes

[1] Deutscher, G. (2010). Through the language glass: Why the world looks different in other languages. Metropolitan books.

[2] From a letter to his editor cited in P. Lee (1996), The Whorf Theory Complex (John Benjamins) p153.

[3] E. Sapir, Selected Writings, (edited by D. G. Mandelbaum, U. Calif. Press, 1949), p. 162.

[4] B. L. Whorf, 1956, Language, Thought, and Reality (MIT Press) p214.

[5] N. Evans 2009, Dying Words (Wiley-Blackwell) p155; the wording is adapted by Evans from literary critic Harold Bloom: ‘We read because we cannot know enough people’.

[6] S. C. Levinson, 2003, in Language and Mind, edited by D. Gentner and S. Goldin-Meadow (MIT Press), p25.

[7] P. Lee (1996), The Whorf Theory Complex (John Benjamins), p14.

[8] Deutscher, p21.

[9] Deutscher, p131.

[10] Deutscher, p148.

[11] Deutscher, p149.

[12] S. Pinker, 1994, The Language Instinct (William Morrow & Co.), p57.

[13] J. A. Fodor, 1985, ‘Précis of The Modularity of Mind’ (Behavioral and Brain Sciences, vol. 8, pp1-42).

[14] J. W. Schooler & T. Y. Engstler-Schooler, 1990. ‘Verbal overshadowing of visual memories: some things are better left unsaid’ (Cognitive Psychology 22: 36–71), p66.

[15] Deutscher, p136

[16] Slobin, D. I. (1991). Learning to think for speaking: Native language, cognition, and rhetorical style. Pragmatics, 1, 7-26.

[17] P. Li & L. Gleitman, 2002 ‘Turning the tables: language and spatial reasoning (Cognition vol. 83, pp265–294), p266.

7 Comments

  • comment-avatar
    Olivier Morin 5 February 2011 (19:28)

    First, thank you for honoring our blog with this brilliant review. You conclude by daring the Whorf-skeptics: [quote] “If you think the effects of language on thought are trivial, try telling that to the ghost of Tim Cole.” [/quote] Let me try and answer the challenge. “Dear Tim Cole – I know your life was ruined by a dreadful miscarriage of justice, for which you have every right to blame the unreliablitiy of eyewitnesses and the tragic sloppiness of justice. But should you blame Michelle Malin’s error on the English language? I very much doubt it. As Nick brilliantly shows, language does affect thought — and this is not trivial for all the intellectuals who see this as an impossibility. But as far as you, your judges, Michelle Mallin and other ordinary citizens are concerned, linguistic framing of cognition is probably not powerful enough to matter a great deal. Could Michelle Mallin’s error have been provoked by the ‘verbal overshadowing’ effect? In theory, it could. But memory research describe many, many mechanisms, more powerful and more frequent, to explain the unreliability of eyewitnesses. Interesting as it is, verbal overshadowing is a weak effect: in strictly controlled conditions (quite unlike the situation of Ms. Mallin), it yields small effects (actually, the effects have been getting smaller and smaller with each new test of the effect). Intellectuals like Nick and I usually care about the existence of these effects – we are often heard to say that if it is ‘statistically significant’, hence real, nothing else matters. But you need to know how strong the effect was, how likely it is to have curbed Ms. Mallin’s thoughts, and our answer must be: as far as we know, it probably didn’t. Suppose that verbal overshadowing was indeed responsible for your plight. Could we then put the blame on the English language? No. First, it is unclear what role language plays in this effect – we do not know exactly how verbal ‘verbal overshadowing’ really is. When Michelle Mallin redescribed the agression, she was not just attaching words to a memory – she was re-experiencing it, trying to make sense of many details, to fixate uncertain elements. She was doing this to tell a story, couched in language, but we do not know to what extent language guided or constrained her recollection. There are some reasons to think that memory can be manipulative on its own. Note that until now, we have been talking of Language in an abstract way. Of course, what matters for Nick’s point is not Language, but English as opposed to Lao or Spanish. Could Michelle Mallin have acted differently if she had spoken Spanish or Lao? Again, there is no theoretical impossibility, but the kind of mechanisms that could have pushed her in the wrong direction, though theoretically interesting, are much too weak to make a big difference in the real world. Today’s neo-Whorfians (people who claim that “the language we speak shapes the way we think”) rest their case on small-but-significant experimental effects showing that, in some restricted areas, language can subtly tinge our thoughts and prime our actions. Nothing [i]big[/i]. You probably don’t care to know how this debate started, but let me just to give you an idea of the stakes. Yesterday’s Whorfians used to claim linguistic influences on thought were strong, so strong they could compel you to see the world in certain ways or significantly block your access to certain ideas. Today’s neo-Whorfians are much more cautious. Still, they seem to agree with Nick that the weakness of these mechanisms and the small size of these effects are no cause for concern. Big or small, Nick writes, “An effect of language on thought is an effect of language on thought.” In other words, they want to make this a qualitative dispute, not a quantitative one. But what got people excited in the first place [i]was[/i] a quantitative issue – [i]how much[/i] does the language we speak constrain us? [i]How much[/i] did it constrain Michelle Mallin?” So, Nick, if you think the impact of Whorfian effects matters less than their mere existence, well… I suggest you try telling that to the ghost of Tim Cole.

  • comment-avatar
    Nick Enfield 6 February 2011 (20:59)

    Thanks Olivier for your engaging comments. I agree that ‘Whorfian’ work needs to be clear about whether the claim being pursued is a qualitative or a quantitative one. For now, I just want to clarify something about my use of the Timothy Cole case as an example. In my post, I discuss the logic of how verbal overshadowing could in principle be involved in linguistic relativity effects. (This idea was pointed out to me by my colleague Asifa Majid; with her and other colleagues at MPI Nijmegen we are exploring it in comparative linguistic work.) The Schooler and Engstler-Schooler study looked specifically at witness identification, and I wanted to cite a real case that would give the idea some impact and concreteness for the reader. In the Tim Cole case in particular, it is clear that there were multiple factors leading to his misidentification, and while Michele Mallin had indeed verbally described her attacker, we can never know whether, and if so how much, this had an effect. Details of the misidentification, and the many factors that contributed, can be found on the ‘Innocence Project’ website (an amazing, and very sobering site, by the way): http://www.innocenceproject.org/Content/Timothy_Cole.php The use of the Tim Cole example in my post was not intended to claim that language, or the English language, was known to be a factor in that particular case. Still, I stand by the argument that language could be a key factor in witness misidentification, and, in turn, in wrongful imprisonment. Such outcomes are clearly very far from trivial.

  • comment-avatar
    José-Luis Guijarro 7 February 2011 (19:03)

    Although traditional, the debate on language shaping thought or viceversa is getting more and more refined everyday. The above discussion is a very brilliant and interesting interchange of views in that topic, to say the least. However, from my point of view, the issue is somehow confusing by what I think is the poverty of the English vocabulary in relation to the concept LANGUAGE. In Spanish, we have three different terms, LENGUAJE, LENGUA and IDIOMA, which would warrant, I think, using those three lexical pointers to signal different conceptual spaces –concepts that seem at a surface level impossible to consider when using the English language, although some researchers (i.e., Chomsky &al in debating the origin of language with Pinker and Jackendoff) do realize that some people might not be looking at the same object when talking about it in one way or another. For historical reasons, though, Spanish linguists, as far as I am aware, seem to treat the three Spanish terms as if they were almost synonymous, like the English or German tradition has it. It seems weird to me that such a case of triple synonymity happens in my native language, so I have tried to adequate my use of these three terminal pointers to different concepts. Following Sperber & Wilson idea in the fourth chapter of their book on Relevance, I use the pointer LENGUAJE (I will call it L1 hereafter) to signal the mental device that allows us (and perhaps other species) to formalize indices of the World around us in order to store and manipulate them mentally in many different ways. As such, L1, for me, is the other side of the coin “cognition/language” (and not, as it is normally conceived, communication/language). Therefore, to ask whether language shapes thought is somewhat redundant. L1 is thought (i.e., it allows creation and manipulation of mental representations). It is what, in another type of World, language does in the innards of computers. I use the term LENGUA (henceforth L2) as the special functional configuration of principles and parameters (to adopt a chomskyan terminology) which acts as a sieve through which linguistic data are filtered in order to facilitate the acquisition of a linguistic communicative tool. I call this linguistic communicative tool, IDIOMA (or L3) and I think it may well be construed along the possibilities of the innate device of L2 with all the important or salient, or whatever you may want to call it, biases that a given environment has forced a human group to formalise in its L1 throughout its existence. It seems clear to me that my non-unitary view of LANGUAGE could, as if by magic, solve the present problem of which shapes which by making it dissolve in thin air. Its effect would be similar to the one that makes a fist disappear just by opening the hand. The problem does no longer exist; it was all a question of wording it differently. Small wonder that when reading debates on that issue, as the one above, every interlocutor seems to be right more often than not!

  • comment-avatar
    Benson Saler 8 February 2011 (15:33)

    First, I endorse Morin’s description of Enfield’s blog as a “brilliant review.” It is that, and I learned much from reading it, and have much to think about in consequence of its thoughtful and lucid discussion of important issues. Second, and with respect to Jose-Luis Guijarro’s comments: That language is a tool for expressing thoughts — and perhaps to some extent for shaping them — is all the more evident when language is employed pejoratively. Numbers of native Spanish speakers in Central America and South America, for instance, reserve the term idioma for their language, for other European languages, and, in places, for still other languages. Such persons, however, tend to use the term lengua when referring to Amerindian languages. They sometimes maintain, moreover, that lenguas — Quiche and Wayunaiki in my personal field experiences — are deficient or limited languages, so that they cannot be used to count above some fairly low number, or cannot be used to frame profound observations or moral abstractions, or something of the sort. The term lenguaje sometimes has an even more pejorative application. As a diminutive, it may be coupled with another diminutive, “inditos,” suggesting a sort of affection, but one subsumed in a patronizing attitude. I note the above without further comment.

  • comment-avatar
    José-Luis Guijarro 8 February 2011 (18:42)

    Does that mean that when you (and by you, I mean YOU) use those terms you point to those objects? I don’t mind, provided that you make it explicit at the beginning (of a debate). The real issue in dealing with abstract or unclear notions is to attain the OBSERVATIONAL level of adequacy. I have tried to show what my (and by my, I mean MY) level of adequacy is in this particular case. I have no pretension to make other people adopt my distinction. What I do attempt is to point to different conceptual places in the hope that I may clarify an issue that otherwise, brilliantly or not, remains a bit confusing, especially with the BRILLIANT reasoning of two very informed interlocutors like Nick and Olivier.

  • comment-avatar
    José Luis Berbeira Gardón 20 February 2011 (06:38)

    My thanks to José Luis Guijarro for calling my attention to Nick Enfield’s illuminating discussion and the ensuing debate. I am particularly interested in Nick Enfield’s concern for decision-making processes. As he states, the strategy of weighing up all the potential alternatives when, for example, deciding what to order at a restaurant by studying the entire menu is implausible, as it would be cognitively costly. In fact, as many scholars in different fields of research have highlighted, in decision-making processes under uncertainty, reasoning alone is strategically insufficient as, in a given situation (such as the restaurant example), it cannot provide us with a way to stop the process of reflection through the selection of one amongst a set of potential alternatives. As Nick Enfield points out, the best way out in making the best decision is to look for ways to narrow the search so as to select the first alternative that provides us with the optimal benefit for lowest cost (in Sperber and Wilson’s terminology, searching for the optimally relevant alternative). In this sense, perhaps, Ronald de Sousa’s (1987: 196) notion of ‘emotions’ as compensating the insufficiencies of reason by controlling ‘salience’ (and, thus, saving us the effort of having to weigh the confirming and disconfirming evidence for every alternative) would be a good way to account for how this search is carried out. As I have argued elsewhere, there are great benefits to be gained from exploring the role of emotional biases in both decision-making and context selection. With regard to the debate on whether language shapes thought or viceversa, I agree with those holding the view that the traditional conception of conceptual categories as fixed entities is misleading, since it misses the fact that there are more concepts in the mind than words in the public lexicon (Sperber & Wilson, 1997). (In this particular respect, I have to admit that I haven’t been able to grasp Nick Enfield’s actual views on concepts and categorization). Research in both cognitive linguistics and in relevance-theoretic lexical pragmatics provides evidence for the idea that most concepts are not specified/lexicalized in the lexicon, but are construed on-line in actual situations of use. In my view, this should be taken as evidence against the idea that language shapes thought. In using language, we have a finite stock of words with a linguistically encoded concept that we have to inferentially adjust in order to be able to convey those unlexicalized concepts that we want to communicate. This suggests that it is thought that shapes language rather than viceversa.

  • comment-avatar
    José Luis Berbeira Gardón 20 February 2011 (11:18)

    When I said that “it is thought that shapes language” I meant “it is thought that shapes/constraints THE WAY WE USE language”.