Dark Matter of the Mind

by Daniel L. Everett

  Knowledge of language is not merely a knowledge of all sentences, but the structure of that knowledge—for example, the mapping between sentences. Another way of thinking about this (one that I would strongly urge) is that language is not a set of any units, such as sentences or words, but rather an understanding of storytelling—how to tell stories, what they should be about, how they should be structured, and so on. Beyond this, language is knowledge of the social group, of the culture from its narrowest community boundaries to its widest extent. This is clearly reminiscent of Bakhtin’s (1982, 284) view: “Any understanding of live speech, a live utterance, is inherently responsive . . . Any utterance is a link in the chain of communication.” According to Bakhtin’s concepts of heteroglossia and polyphony, everything we say is a mixture of things we have heard others say and the interaction with them in dialogue.

  Translation, like culture itself, is a dance, an action-reaction pairing in which two beings adjust themselves according to the necessity of the environment—for example, maintaining the same speed while dancing. Thus not everything—perhaps not much at all, even—in human interactions is the result of mental representation, but is rather acting together. This brings us to a deeper and more foundational question we have avoided so far; namely, whether translation is even possible. It is a truism among linguists that all languages are capable of expressing the same things. But if language is formed symbiotically with culture, does it necessarily follow that translation of anything from one language to any other language can be done? What does translation entail?

  No Universal Intertranslatability

  Consider the possibility of translating the following sentence into Pirahã (thanks to Geoffrey K. Pullum [pers. comm.] for suggesting this example):

  My cousin totally ran out of cash during his second year at art college in California and had to draw $10K from the Bank of Mom and Dad to pay his tuition.

  What does this sentence show? First, it shows a syntactic device common in English that is lacking in Pirahã: embedding. This means that in Pirahã, this single English sentence would have to be expressed by more than one sentence, since without embedding one sentence cannot be placed inside another. (The absence of embedding in Pirahã results from a lack of recursion in Pirahã; D. Everett 2014b; Futrell et al., forthcoming). The absence of embedding need not be fatal to translating the content, though it does demonstrate that translation from English to Pirahã cannot be done by matching forms, thus losing some of the translation’s accuracy (according to Sontag, though not necessarily so according to the dynamic translation philosophy).

  Neither, however, can a translation match the English concepts with Pirahã concepts. There are no words in Pirahã for mother or father, only a generic word for parent, which, as we have seen, has a wider meaning than biological parent, referring also to someone one is hoping to get something from, or someone who holds power over you in a particular situation. There are no quantifier words, hence “totally” cannot be translated. Further, Pirahã culture has no money, though they do have a vague concept of it from watching Brazilians. They have no numbers, so one dollar or ten thousand dollars would not be translatable. They have no concept or word for art. They have no word for year, per se, though they could say “water,” referring to a high-water/low-water cycle (one year). They have no concept or word for bank. Thus this entire sentence is ineffable in Pirahã, in both form and content. Therefore it is not possible in practice to translate from all languages to all languages, QED. It might be possible, though, if speakers and hearers were to come to share sufficient dark matter, via prolonged cultural contact, that they became bicultural. But that is not the case here, thus underscoring the importance not only of dark matter to translation, but of translation as a test for degree of dark matter overlap, of knowledge of a language, of cultural understanding, and so on.

  Whether translation is actually possible in a complete or interesting sense depends on the objectives of the translation. A linguist translating a language for other linguists will have different objectives than a missionary translating the Bible or a company translating an operating manual. The important point is that translation is a cultural activity, putting one culture into terms of another culture, not just transferring one language into another. Language is only part of the translation process. This is shown by the texts we examined in earlier chapters and by the example above. The implications of this are profound: not all things can be said in all languages. This is the implication of the notion of dark matter in grammar and cognition we are considering here.

  Another way in which translation interacts with dark matter revolves around the notion of relevance. Dark matter shares in common with Sperber and Wilson’s (1995) relevance theory the idea that we need assistance to process the vast amount of input the world bombards us with. Relevance theory says that we attend to what is relevant, where relevance is determined by context, topic, interlocutors’ assumptions, culture, and so on.

  The idea of relevance is motivated by a “bottleneck” theory of cognition—namely, that since our attentional resources are limited, we use some principle of relevance to license our disregard for the majority of the input and to place our focus on a narrow band of the environmental stimulus. (Broadbent [1958], Treisman [1991], and others made the point that we can switch our attention to previously unattended information if that becomes relevant, such as overhearing our name in an across-the-room conversation, “sensing” someone staring at us, seeing flames, being attacked, or seeing the main speaker walk on stage.)

  Sperber and Wilson’s work was inspired by the writings of philosopher Paul Grice (1991) on maxims of conversation, constituents of the larger cooperative principle. This principle describes our linguistic exchanges as making our contribution, such as is required or desired at a particular stage in conversation by the accepted purpose or direction of the talk exchange we are engaged in.

  The four conversational maxims that manifest the cooperative principle are as follows.

  1. MAXIM OF QUALITY

  Do not say what you believe to be false. Do not say that for which you lack evidence.

  2. MAXIM OF QUANTITY.

  Make your contribution no more nor less information than is required.

  For example:

  Q: “How are you?”

  A1: “Uh-huh.” (underinformative)

  A2: “I was constipated yesterday and not sure if I should eat more cheese today, given that my lactose intolerance isn’t going away . . .” (overinformative)

  3. MAXIM OF RELATION

  Be relevant.

  This is harder to control for and study, because culture is the foundation of relevancy. The hearer will assume that the speaker is saying something relevant, and the speaker will assume that the hearer will look for the relevance; shared cultural values are crucial. The relevancy detected/assumed/constructed need not be literally there, since the culture-specific knowledge and cultural nature of the exchange will fill in the gaps. For example:

  Q: “Have your parents arrived?”

  A: “Mother is ill.”

  The hearer will infer that the mother’s illness is relevant to the question of whether the parents have arrived. If one speaker asserts in this context that his mother is ill, then since this must be relevant to the question, and the parents must not have arrived. Or:

  Q: “Have your parents arrived?”

  A: “There were fish in the river.”

  This is a stretch. How could fish in a river have anything to do with whether one’s parents arrived? They do not, literally. But the hearer will again assume that there is something relevant in the answer to the question and so think along the lines of “Perhaps there is a river his parents needed to drive over with a bridge of some sort that cannot be employed when there is also a high concentration of fish in that section of the river.”

  A conversation is another form of dance. Each dancer assumes that his or her partner is trying to dance the same dance. Relevanc
e and cooperation, however, are culturally influenced and individually determined (Henrich and Henrich 2007). There is relevance in the sense of the culture as we are using that term here, as well as relevance to each partner as self-constructed, and thus what they each choose to attend to at any particular time, how they choose to interact, and so on.

  MAXIM OF MANNER

  Be clear, avoid obscurity, avoid ambiguity, and say no more than is required.

  These are not grammar rules. They are expected by the hearer as means to more easily interpret what the speaker means. For example, if someone asks you, “Do you believe that abortion should be illegal?” the expected answer is “yes” or “no.” Thus if, as candidates do in political debates, you begin your answer with “My campaign has always been clear that some issues are complex,” then the hearer will take this as a poor answer, one that for some reason flouts the maxim. (All the maxims may be flouted.) This doesn’t mean that the offending answer is meaningless, just that the speaker intends to say something other than what was expected.

  Translation is also an important test of the understanding of dark matter because the ability or inability to translate tells us about shared concepts. Concepts are crucial to our hermeneutics and translations. They are often held up as the ultimate example of knowing-that vs. knowing-how. One of the most influential and original discussions of concepts in recent years is found in the work of Susan Carey, in particular her 2009 book, The Origin of Concepts. Her theses in this book are (i) that humans are born with rich conceptual representational structures (back to Bastian); (ii) that human perceptual abilities help us to identify innate concepts; (iii) that our innate representational structures/systems never change, and therefore, there will be cross-cultural conceptual universals; and finally, (iv) that we have mechanisms for transcending core cognition. We can create culturally influenced representations that are incompatible with the fixed set we are born with.

  Carey’s theses contrast with Brandom’s. Again, Brandom, in all his recent work, argues that concepts are not given but “earned,” by deriving, demonstrating, and learning their meanings via inference. And only Brandom ultimately provides a satisfying account of shared concepts—concepts built by interaction and individual apperception, varying tremendously from culture to culture and family to family.

  The theory of dark matter developed throughout the different divisions of the preceding discussion takes a skeptical view of nativist perspectives of concepts (e.g., Carey’s). Some values and emotional centers of our brain have a biological source, though the expressions and conceptualizations of these—the value “avoid pain,” “avoid death,” “be happy,” and others—have local cultural interpretations.

  Before innate concepts are proposed, we must survey world cultures, account for differences, assess proposals of universals, and understand the range and typology of learned concepts, for even Carey’s theory of concepts recognizes that many concepts (e.g., “US president”) must be learned. However, once we have understood how, why, and which concepts are learned transculturally, what is left for nativism, aside from standard poverty of stimulus arguments? And what, after all, does “poverty of stimulus” mean in practice, other than that we cannot think of a stimulus responsible for a particular concept, action, or other learning? As many have said in the past, when looked at carefully, the expression “poverty of stimulus” is interchangeable in practice with “poverty of imagination.”

  Problems in Translation

  A chapter on translation should perhaps include more examples of the difficulties of what Quine (1960) referred to as “radical translation.” Recall that Quine’s discussion of radical translation and the “indeterminacy” of translation flow from Quine’s concern with a linguist confronting a linguistic community that speaks a language unrelated to any that the linguist is familiar with—what Pike called a “monolingual setting.” Quine (based on my reading of Quine 1985) came up with the example, I believe, via his contact with Kenneth L. Pike (my first linguistics professor). The idea is roughly that we have no way to exactly know what the referents of a given term are. This section, then, is an expansion on the earlier question of intertranslatability.

  One of the first questions visitor to the Pirahãs ask me is “How do you say ‘hello’ in Pirahã?” Or “How do I say ‘thank you’?” These were once my questions too. How does one greet others among the Pirahãs? Or take leave, express gratitude, and so on? Well the answer is simple—generally one does not do these things. Such uses of language are called “phatic” language, and Pirahã, among other groups, simply lacks this type of language. One can say “I am going now” for “Good-bye” or “You gave me this” rather than “Thank you,” and so on. But normally one doesn’t say anything at all. One leaves. One arrives. One accepts. One gives. No special declaration is needed culturally—after all, these actions are obvious. This means, however, that there is no way to accurately translate thanks, bye, hi, and so on, into Pirahã. So as I once translated portions of the New Testament into Pirahã, I might translate thank you as “I accept this. I will pay you back.” Yet, though Pirahãs might say this kind of thing, they simply use no formulaic phatic language. This is a translation problem of “category vs. no category.”

  Another example of category problems is what I call the partially matching category. When I first attempted to translate some Pirahã words into English, I used what to my then Christian mind were the closest English equivalents. This confused my readers and, ultimately, the Pirahãs as they received my attempts to translate. For example, one of the hardest terms for me to understand was kaoáíbógí, which literally means “fast mouth.”

  In the evenings, sometimes throughout the night, one can hear a falsetto male voice coming from the jungle into the village. It may give advice on where to fish, what foreigners to avoid, or how to spend one’s next day, among other things. It can talk about what living underground is like for the dead. It may tell stories about jaguars. Or the speaker may walk into the village naked and start acting vulgar and saying vulgar things. The people will say, animatedly, that “a fast mouth is here.” And when I reply, “But that looks just like so-and-so,” they respond, “Yes, fast mouths look just like Pirahãs. But you can see that they are not Pirahãs: They are naked. Pirahãs are not. They talk high [in a falsetto]. Pirahãs do not talk high.”

  Kaoaibogis can live under the ground, under the water, inside tree trunks, everywhere. They do not have blood. Thus I originally interpreted these jungle entities as “spirits.” But this is incorrect. First, kaoaibogis are material. Second, they play no role in a larger (e.g., religious), belief system—no more than a jaguar or a fish does, in any case. Their function in Pirahã society is for drama, for humor, for scaring others, for correction of children, for advice. They are real and physical. They are similar to people but not identical to people. They are similar to people lacking normal cultural prohibitions. Beliefs about them—what they do; how they live; whether they are good, bad, or neither—vary among the Pirahãs. Yet there is no sense or referent of kaoáíbógí that makes it translatable as “spirit” in English. They are entities that do not map well into a North American ontology or the English language.

  Kaoaibogis do not map neatly to fact or fiction. The Pirahãs enjoy them and see them, but they are liminal creatures, neither mundane nor otherworldly. They can be seen, heard, smelled, touched, and so on. Yet they also are funny, even when trying to be serious, as though the people were embarrassed by the person “playing with the jungle entity.” At the same time, all Pirahãs swear that kaoaibogis are real. The person they resemble (what a nonbeliever might refer to as the “actor”) will refuse to admit that he (never a woman, even when the kaoaibogi dresses like, talks like, and is claimed to be female) was present when the kaoaibogi was present and will add that he never saw the kaoaibogi.

  The “fast mouth” label is just one for a series of nonhuman, humanoid entities. There are also “big tooth” (xaitoii) and “jaguar�
� (kagaihiai), among others. A kagaihiai is not a real jaguar, it just looks like (and can act, kill, and eat like, etc.) one. And so on. These creatures are all just among the multitude of types of jungle entities encountered by the Pirahãs, and like other entities, each individual may have a personal name. It isn’t clear to me whether they change individual names as the Pirahãs do, but I suspect that they do.

  With this background, we see that kaoáíbógí has/have no simple one-word or one-phrase translation into English. These entities are neither fact nor fiction in our sense, neither friend nor foe. In a sense, to quote Philippe Descola (2013), they are “beyond nature and culture.” We ought not to translate these terms as spirits, sprites, demons, possessed people, or any other off-the-shelf English word. In fact, the point is not that there is no one-word translation, but that there is no translation, period. At best we can describe them, but the Pirahãs’ conception of and classification of sentient entities in their world fails to match our own (and vice versa, of course). Likewise, the Pirahãs have no concept of “god” and thus cannot, without much learning about, say, Brazilian culture, even approach an accurate understanding of what we mean by a supreme deity. Previous missionaries tried to translate God as baixi hiooxio, “up-high father.” I also attempted to use this as a translation for God. But the term, like the concept it is trying to communicate, turns out to have no purchase among the Pirahãs. They have no supreme being, no creator, nor any savior among their beliefs about the world.