Beyond Utrecht

The universe has four dimensions. Language is not so big, but perhaps it can boast two dimensions.

Download PDF of post

Two years from now the next Evolang conference is scheduled to be held in Kyoto, Japan and I’m guessing that by then we will have a clear answer to the close of last week’s post on this blog, “We’ve got to understand how biology and culture are two dimensions shaping every utterance.” I’m optimistic because scholars already know what each dimension is about, and it looks like some people have identified the glue that holds the two dimensions together. So let’s hop into a time machine and see what understanding is likely to be clear when the Kyoto conference finishes up.

Most linguists and other scientists contributing to this field are primarily concerned with language as we know it today. It is a perfectly reasonable position, and what a strange world we would live in if that focus was not the case, but that interest forces scholars to consider a wide variety of topics that this blog, with its hard focus on evolutionary origins, can ignore. Language’s core topic is the concrete world, present or imagined. Language today has many more topics than that. Metaphors and symbols make abstract and invisible topics discussable. But metaphorical and symbolic language builds on the linguistic tools evolved to talk about the physical world.

1 — The Biological Dimension

We talk about the physical world by drawing attention to it and specifying it perceptually. Take the opening sentence of Peter Matthiessen’s masterpiece Shadow Country, “Sea birds are aloft again, a tattered few.” It is a kind of poetry, but it is also physical. The one questionable word there might be tattered which we ordinarily associate with clothing, but it is used to describe the physical condition of something physical (birds) and I think it passes as the outer edge of the kind of language I’m talking about. Maybe a Homo erectus poet or two pushed language this far, especially when you notice that a tattered few has not been woven syntactically into the sentence.

More complex, but still completely physical sentences can be found in the book—e.g., “While his kin said good-bye to their guests, Lucius inspected the framed photos of Billy Collins and his sons.”

A sentence like that puts the whole scene directly in front of the reader. Every word is about the physical world. Jean-Louis Dessalles points out that a complex sentence like this one expresses a complete gestalt, with foreground (Lucius inspected the framed photos of Billy Collins and his sons) and background (While his kin said good-bye to their guests). Matthieson could have handled it differently, as he did in the first sentence, keeping each part separate and equal. It is the first word, while, that organizes the sentence into one, whole perception. Language that expresses only concrete things had come as far as it could when it became able to unite foreground and background into one perception.

Some time back I posted a note that pointed out that the elements of language that generative grammarians call arbitrary can be understood as basic features of  perception. (see: The One Model that Works) These features are:

• point of view (aspectual reference): the kin said good-bye to the guests could just as easily be the guests said good-bye to the kin and the while could be moved to the other clause, switching what’s in the foreground. As it stands the sentence tells us know what matters to the storyteller, and we see it the same way.
• perspective (edge features) works like a zoom lens enabling the teller to move in or out on a scene; the whole dependent clause (while …) could be removed without losing the sentence’s core meaning. Also the details of the photos could be removed. The sentence would be much weaker if it merely said Lucius examined some photos, but language offers the producer the choice of how far to zoom in or out.
• focus (headedness), the language is about something; attention gets directed to many points in this sentence (kin, guests, Lucius, photos), and each word in the sentence relates to one of these foci.

Every natural language has these features because it rests on the biology of perception. No matter what language a child is destined to speak, these things will be there. This dimension is not syntactical or even semantic. It is more on the level of a governor, the thing that determines the syntax and semantics appropriate to the foci.

2 — The Cultural Dimension

The variable rules for expressing the governing perception encompass all the semantics and syntax of a particular language. The elements of perception can be distinguished and categorized in an enormous number of ways, and no one way has claim to superiority over another. It is important, however, that all members of a group be able to share their perceptions, so they must distinguish and categorize perceptions in a settled way. The rules are arbitrary, but their drift is constrained by the need to retain meaning; the governing perception must be discoverable.

 It is a mistake to expect universals in this dimension. The semantics and syntax of any statement reflects the evolution and drift of a particular language and culture. Generalizations about the whole of this dimension—e.g., syntax is structured hierarchically—reflect either the biological dimension or the glue that holds the two dimensions together.

3 — The Glue Linking the Dimensions

The basic glue holding the biological and cultural dimensions together is the power of attention, one of the most mysterious functions of the brain. Every utterance that is not just a high-powered animal signal—e.g., halt; sure-sure—uses attention to link statement and governing perception. The speaker attends something and utters the apt word(s) of a particular language. The listener hears the word(s) and pilots attention appropriately.

A second glue uses another function of perception, the gestalt. That’s the ability to bind separate elements of a perception into a coherent whole. Examples of gestalts include the way auditory perception turns musical notes into a melody or the way visual perception constructs motion out of a series of still pictures. A sentence like, “His kin said good-bye to their guests,” has two separate foci, kin and guests, but we can imagine the scene as a whole. That’s a gestalt; the construction is in our head, not in the sentence structure. Change the word said to inflated and the gestalt fails. The structure is unchanged, but the sentence has become meaningless. A meaningless remark is one that does not form a gestalt.

The job of an editor is to ensure that the glue in any particular sentence binds the words into a whole. Is attention piloted successfully, or does the reader get lost? Does the complete sentence come together in a gestalt or is it too much to ask of the audience? Sentences may pass syntactical scrutiny but still not be intelligible.

4 — Still Unexplained

An obvious problem is that these days we talk about plenty of things that are not perceptible. Derek Bickerton  offers a good sentence to challenge the ideas in this post: My trust in you has been shattered forever by your unfaithfulness. We’ve got both abstract concepts (trust, forever, unfaithfulness) and a metaphor (shattered) in this sentence. Where did they come from? The short answer: I don’t know. The long answer: Although I do not know, I am not especially worried because I believe this account gets us through the fixing of language in the species; we began talking about perception and many generations later came imperceptible topics.

I base my confidence on the way abstract sentences are structured as though they were describing perceptions. Bickerton’s sentence is written as though it is about something physical: My house in Alabama has been shattered to pieces by your bomb. The structure of the sentence organizes the phenomenon as a real world action complete with possession (my, your), subject (unfaithfulness), object (trust), indirect object (you). These relationships are brought together by a metaphorical verb. This kind of expression reflects a language and way of thinking that is much more attached to the concrete than to the abstract symbol processing of a modern computer.

Frankly, I don’t know that humans of a hundred thousand or more years ago engaged in any abstract reasoning. Their thinking seems to have been much more mythological, which is concrete or metaphorical or mysterious. Mystery symbols that express an identity or power without providing something concrete beyond the symbol itself seem pretty certain to have existed by a hundred thousand years ago. All of these things constitute more than straight perceptual thinking, but seem to be dependent on them. Maybe we will be able to take up these mysteries after Kyoto.

Comments

[Paul Strand]

The most striking recent advance in the study of language is not the empirical support against the necessity of language-specific structures or innate information for language production (see Newman et al., 2010 for the most recent neuroscience evidence; See Moerk, 1989 for the psychological evidence). After all, the support for that idea was always more philosophical than empirical (Chomsky, 1959). What is perhaps most stunning is the development of procedures for teaching language to humans who, because of neuro-developmental disorders such as autism, were previously thought to be incapable of it (Greer, 2008).

What is becoming clear from these efforts is that the dissection of language into functional rather than structural parts is necessary to achieve these successes. These studies reveal that language learning is intimately tied to its usefulness to the organism learning it (see any issue of The Analysis of Verbal Behavior at pubmed). Moreover, what constitutes an environment that is “rich” for language learning is different for different people.

In light of the Utrecht conference and recent neuroscience findings, perhaps the search for understanding the evolution of language might take as its starting point how it is that non-verbal repertoires are replaced by verbal ones. At the level of organisms, this happens when the verbal repertoire (however simple) is more effective than the prevailing non-verbal one. For this reason, the first step in teaching language skills to individuals does not involve teaching them to name things (nouns, verbs). Instead, the first steps usually involve teaching them to make requests. (This gateway skill constitutes a basic unit for functional but not structural accounts of language). Once language-less children can do this, they often can be taught to engage in naming, sentence construction, and all sorts of other verbal operations (see Hayes et al., 2001; Greer, 2008). Therefore, it may make sense to consider all language as an elaboration on the process of making requests. How this elaboration takes place might make for a fruitful line of inquiry.

References
Greer, D. (2007). Verbal behavior analysis. Allyn and Bacon.
Hayes et al. (2001). Relational frame theory. Kluwer.
Moerk E.L. (1989). The LAD was a lady and the tasks were III-defined. Developmental Review, 9, 21-57.
Newman, T. Supalla, P. Hauser, E. L. Newport, D. Bavelier. Dissociating neural subsystems for grammar by contrasting word order and inflection. Proceedings of the National Academy of Sciences, 2010; 107 (16): 7539 DOI: 10.1073/pnas.1003174107
-----------------------
BLOGGER: Paul's posts are always pretty interesting. Thanks.

[JanetK]

I see the biological dimension as somewhat more extensive than outlined here. For example I think the root of what I would call 'the basic vocabulary' is biological. We would be surprised to find a language that did not have a word for 'arm'. This is probably because an arm is an object with 'natural boundaries'. In order to make perceptual sense of our own movements and those of others, our brains create objects and those objects have natural boundaries. So before we have a word for arm, we have a concept of arm because we need such a concept for rapid and accurate perception and movement. So I would guess that the ability to form concepts would pre-date the ability to use words and would result primarily from the biological ability to perceive and act. It is hard to conceive of a brain that works without breaking the world into objects and those objects are formed by using the natural boundaries of surfaces and discreet movement. Those objects have to be stored somewhere in the brain for reference and are therefore concepts of a type. As soon as the concepts are used for communication they will be become words.
I like the idea you have put forward many times, that words guide joint attention. But I think the joint attention is not so much focused directly on the world but is focused on shared concepts. Pointing is all that is needed to guide attention to something directly. Words are needed to guide attention to concepts.

[Adrian Morgan]

I do wince a bit at the choice of "dimension" as the word for describing the two aspects of language.

A characteristic of dimensions is that it doesn't matter how you combine them. Three paces east plus five paces south plus another three paces east is the same as six paces east plus five paces south is the same as five paces south plus six paces east.

Used metaphorically, "dimension" may be a suitable word to describe aspects that are essentially independent and can be described without reference to each other, but if you're talking about aspects that influence each other in complex ways (or have the potential to do so), then I think the word is better avoided.
--------------------
BLOGGER:I don't want to make you wince, but I did mean to use the term "dimension" fairly seriously. It's not like length and width, more like space and time. You can't run around at will in the temporal dimension either. But the two are glued together via the speed of light. That's the quality I was trying to get at. The biological and cultural dimensions can be discussed independently, but at any moment both dimensions define an utterance. It also has some analytical value because it helps keep the two sorted. What is the most common mistake in linguistic analysis? Treating the biological and cultural dimensions as though they were the same thing.