Defining Language

        John Milton (1608-1674) was a master at using language, but did he know what language is?

Download PDF of Defining Language

One of my most faithful readers and commenters asks (here) what I think of Derek Bickerton’s “notion that imagination, ‘de-localization’ or ‘offline thinking’ are the main characteristics of language.” It’s a worthy question and deserves a serious answer, so I suppose it is time to look at rival definitions of language.

The challenge of defining language is in its variability. Language can make puns, but few of us would accept a definition of language as a tool for punning. On the other hand, can a definition be complete if it does not allow for something as universal as punning? I have not read one that managed to include it. Super-poet and man-of-letters John Milton wrote that language is “the instrument conveying to us things useful to be known.” A century later Samuel Johnson, in the preface to his dictionary, narrowed the definition even further, asserting that language is “the instrument of science [knowledge]” and words are “the signs of ideas.” In his Life of Cowley, Johnson called language “the dress of thought.” Yet, while language can obviously be used to express thoughts, it can do more than that.

One thing Johnson and Milton agreed on was that language was an “instrument,” a tool or means of accomplishing something. That view prevailed until Ferdinand de Saussure narrowed the definition still further. He asked, as a professional linguist, just what is it that linguists study and answered that, “The linguist must take the study of linguistic structure as his primary concern, and relate all other manifestations of language to it.” The reason for recommending this approach was that it was the one part of language that is of no interest to psychologists, literary critics, anthropologists, philologists, etc. For linguists, language became a formal abstraction, a series of units subject to abstract laws about relationships (structure) between its elements. The most common complaint about this view is that the study of meaning went out the window, but also important was the loss of the idea that language has a use.

Here’s the rub that sets me and linguists running off in different directions: abstractions have properties; means have functions. Our definitions set us looking for different sorts of explanations. A linguist looks for the evolution of properties from which language will emerge. I look for the evolution of an organ or behavior which will accomplish the functions of language.

A linguist interested in language origins thus concentrates on finding, among all the properties of language, the primary one that can lead to the emergence of the rest of the properties as well as language itself. Bickerton’s book, Adam’s Tongue, reflects this attitude quite clearly. He first thought syntax was the property without which language cannot exist, but changed his mind. He thought it might be symbolism but came to focus on displacement, the capacity to talk about something outside the here and now, something we can imagine rather than perceive.

Personally, I am quite sympathetic to this idea that imagination is central to being human. My last three books have all been examinations of scientific imagination and how it works. But from the perspective of a investigator into speech origins, I’m dubious. Language can talk about the here and the now, but that is a quibble. More important is that evolution builds on organs or actions not properties. Something working had to appear.

A linguist like Bickerton knows perfectly well that evolution selects organs and actions for their functionality, and part of his anguish has been looking for an abstract property of language that might be selected for its function. He is very aware of the logical difficulty: language is an abstraction whose properties are only useful as part of whole, but evolution does not produce novelties whose complex parts have no value outside the whole. Therefore, language could not evolve.

Bickerton hopes he has found a way out of the conundrum. Whether he has or has not appears to be a matter of taste. Are you persuaded by the argument?  If yes, good; if you have questions, there are neither fossils nor thought experiments to help resolve the matter.

Bickerton’s conundrum disappears if you come at the problem from the other end. Speech is a behavior with many functions, but the functions do not depend on one another in order to exist. They depend on the behavior. I can call to somebody without having syntax, or I can make a grammatical announcement without having symbolism, or I can engage in a marriage ritual without punning.

Bickerton, by the way, knows perfectly that what evolved was a behavior and the behavior was selected for its functionality. But he has been stuck with his linguist’s conundrum. The irony has come full circle. Saussure gave linguists a view of language that ensured that linguists and only linguists would study it, and now they are trapped with a definition that elbows out the contributions of biologists, paleontologists, or anybody else.

I am not a linguist, so my blog pays much more attention to things like vocalization, perception, and attention—things that everyone knows are absolutely essential to speech, but which have no place in the linguist’s formal analysis. In my view, speech emerged from very old capacities that were organized in a new way to give those old abilities a new function. On this blog language is defined as a means of perceiving by other means. It pilots joint attention to produce a shared perception.

And what about displacement, the capacity to speak and think about things that are not present and may not even exist? Sure, that’s a major function of language, although I doubt it was the original one. When I imagine my mother’s face I use the same visual cortex that I used when she was alive and I looked at her face. It’s a neat perceptual trick, but nothing miraculous.

Comments

[JanetK]

This is a very informative post and goes to the heart of why people are talking past one another rather than engaging in real discussion. Stick to your approach!

[raymondw]

John Milton's definition of language: “the instrument conveying to us things useful to be known.” is not very useful. My eyes, tongue, limbs, ears, breath, pain nerves, hands, in fact, nearly any part of my anatomy. are also “the instrument[s] conveying to us things useful to be known.” Samuel Johnson's definition is much too narrow. In discussing Bickerton's ideas you are confusing evolution (i.e., change over time) with one of its mechanisms, natural selection. What does the selecting is the environment in which the organism finds itself. Some changes lead to death, other changes lead to survival and the passing of one's genes on to the next generation, but only if those genes lead to a phenotype compatible with the environment in which it is in. What survival power does imagination have? In what sense can imagination cope with the environment and its contingencies? To survive an organism must adjust to those contingencies. How does imagination serve that purpose? Before you can have imagination don't you first have to have experience?

You're on the right track when you talk of language as behavior and not some kind of abstract structural system. But what kind of ontogeny leads to this behavior? What kind of interactions between the individual organism and its environment lead to language behavior? From observing these interactions and experimental analysis of these interactions can we uncover the operating principles of initiating, maintaining, and varying this behavior? Finally, what differentiates this behavior from all other ways of behaving? Once we understand these things, then maybe we can come up with a satisfactory definition of language.
----------------------
BLOGGER: In defense of Milton, the man would have been very surprised to have the senses or anatomy brought into it. He assumed that the body was a gift from God, while language was Adam's invention.

[Michael]

Studying linguistics myself, I'm not sure if I agree with your description. The structuralist and formalist approaches to language may have been very influential and at times dominant in the 20th century, but there always were competing and equally important strands of linguistics. Karl Bühler for example, who created the "Organon"(Instrument)-Model of language(which he partly derived from Plato, who coined the idea of language as an "instrument")has been very influential in linguistics. He also saw the direction and creation of attention as one of the most cruial aspects of language. Other psychologists and scholars who were very much concerned with the functions of language also influenced linguistics over the years, as can be read in Chris Sinha's thorough historical analysis: http://www.port.ac.uk/departments/academic/psychology/staff/downloads/filetodownload,71770,en.pdf

Cognitive Linguistics is also commited to integrating language into a more general and functional account of cognition, and it is not a coincidence that a major school that competes with formalism is called functionalism. It is called so precisely because it is concerned with the functions performed by language. Roman Jakobson and the 1930s Prague School of linguistics were also heavily influenced by semiotics and de Saussure but still did important work on and were to a large degree concerned with the communicative functions of language.
Bickerton comes from the Chomskyan/Formalist school of linguistics, and many of his shortcomings or his perspective in general can be explained by this. But these views are in no way representative for the whole of linguistics and you and linguists in general do not run in different directions - quite to the contrary.

[Jesús Sanchis]

Defining language is obviously no easy task. Sometimes, I feel that some of the definitions are not about 'language' as we know it, but about the pre-requisites for language. Let's see an excerpt from the post:

"On this blog language is defined as a means of perceiving by other means. It pilots joint attention to produce a shared perception".

Does it make sense to define language without mentioning a community of speakers? The only condition for human language is that there is a human group behind it. Language is born, articulated, constructed as a collective thing. Defining language outside this sphere is a poor attempt at understanding it. Descriptive grammarians, structuralists and Chomsky share the same mistake: they confuse 'language' with a 'metaphor of language', namely that language is basically a set of rules. Not to mention other definitions with partial views of language. On the whole, I think that what is being discussed in this post is not "definitions of language" but "definitions of some aspects connected with language".
------------------
BLOGGER: I was hoping that "shared" implied the community of speakers.

[Mariana]

Excelent jesus, sanchis, I completelly agree with the ending.

[JoseAngel]

Thanks for the ongoing commentary on Bickerton. I found his book very suggestive, but so are your comments and objections as well. As you say, language is a phenomenon with many sides to it, and all the more complex when these different sides seem to have complex effects on the others as they develop, in a feedback process which led to the language explosion. A slo-mo explosion, to be sure, but fast enough compared with the manners of ants. As an instance of this feedback, just think of combining Bickerton's emphasis on de-localization with your own take on joint attention. A process of joint attention to something which is not here surely requires a more developed instrument, and also a more elaborate social capability for joint attention, than interaction about present objects. I find that both your approaches, while they may place an emphasis on different aspects of the whole, are complementary rather than antagonistic.

[Cory]

"Saussure gave linguists a view of language that ensured that linguists and only linguists would study it, and now they are trapped with a definition that elbows out the contributions of biologists, palaeontologists or anybody else."

I agree very much with this statement! There has been a call for a while now that evolutionary linguistics needs more participation from people who know about linguistics, but I think it's equally important that linguists are aware of research in the origins of that which they study! Thank you for the great post.