The Psychology of Communication: Seven Essays

by George Armitage Miller

Amused and amusing in expression, these essays are united by a concern for problems at the intersection of scientific psychology and communication theory.

Genres: Psychology, Nonfiction, Essays

192 pages, Paperback

Published January 1, 1970

Reviews

[Liam Porter]

This short book is a collection of 7 essays by the psychologist and something of a linguist, George Miller. The first 6 are shorter and all concern the Psychological research program "Communication Theory." It's central metaphor is of the telecommunciations system in which information is transmitted and received under the restraint of bandwidth. A bandwidth can only handle so many bits of information in a space of time, and interfering "noise" can also limit the efficient transmition of information. It strook me that Miller is rightfully conscious that this metaphor has limitations and could be misleading, and he is responsible with how he applies it, and takes the picture it paints with a pinch of salt - fully aware that he is employing a tool.

Considered as communication channels, human beings have a rather high noise level. I would like to call your attention to the exact wording I have used. I said that this general law ust hold for human beings insofar as human beings perform the particular function in question. What I have not said is that, since the laws governing communication channels can be applied to human beings, human beings are nothing but communication channels. If you intend to explore further into this suburb of science, you should mark this difference well. Always distrust the man who tells yo uthat human beings are "nothing but" something else, for he is deliberately concealing from you the abstraction on which his claim is based. Human beings are nothing but human beings. p.111

You only need to first the first chapter to find out how obliging Miller is invite severe criticisms of the model.

The first criticism to be made is that when it comes to the human handling of "information," there seems to be no limit to what counts as a "unit," since "associated" concepts may always be "chunked together" (Douglas Hofstadter's phrase). That said, there is a crude approximation to "chunking" in computer science, in which complex operations can be made into shorthand via metalanguages):

Rehearsal or repition has the very important effect of organizing many seperate items into a single unit, thus reducing the load our memory must carry and leaving us free for further thinking. In terms of logic, the process is like the substitution of a single symbol for a longer expression which would be clumsy to write each time we wanted to use it. p.14

And the most significant objection, I would say, would be the mysterious faculty of analogy-forming that can "render" the unknown in terms of the known:

It is my own opinion that man's perculiar gift as a component in a communication system is his ability to discover new ways to transform, or to recode, the imformation which he receives. It seems to me that the very fact of our limited capacity for processing information has made it necessary for us to discover clever ways to abstract the essential features of our universe and to express these features in simple laws that we are capable of comprehending in a single act of thought. We are constantly taking information given in one form and translating it into alternative forms, searching for ways to map a strange, new phenomenon into simpler and more familiar ones. The search is something we call "thinking"; if we are successful, we call it "understanding". It is possible that the development of high-speed digital computers will soon strip even this last shred of dignity from the human mind. In any case, it is quite clear that man is a miserable component in a communication system. He has a narrow bandwidthm a high noise level, is expensive to maintain, and sleeps eight hours out of every twenty-four. Even though we can't eliminate him completely, it is a certainly wise practice to replace him whenever we can. [...] Our society has already made the first steps towards eliminating human bottlenecks from communication systems, and the years ahead are sure to bring many more. p.55

There are some very interesting ideas contained within one short passage, here. I find the notion of "single act of thought" especially intriguing, especially regarding the difference between the first self-conscious pronunciation of an utterance in a foreign language, into the fluid and reflexive use of it at advanced levels:

A popular belief regards human language as a more or less free creation of the human intellect, as if its elements were chosen arbitrarily and could be combined into meaningful utterances by any rules that strike our collective fancy. The assumption is implicit, for example, in Wittgenstein's well-known conception of "the language game". This metaphor, which casts valuable light on many aspects of language, can if followed blindly lead one to think that all linguistic rules are just as arbitrary as, say, the rules of chess or football. As Lenneberg has pointed out, however, it makes a great deal of sense to inquire into the biological basis for language, but very little to ask about the biological foundation of card games. [...] Those who think of language as a free and spontaneous intellectual invention are also likely to believe that any animal with a brain sufficiently large to support a high level of intelligence can aquire a language. This assumption is demonstrably false. [...] nanocephalic dwarfs, with brains half the normal size but grown on the human blueprint, can use language reasonably well, and even mongoloids, not intelligent enough to perform the simplest functions for themselves, can acquire the rudiments. p.86

In a footnote to this, and illustrating historical denial of the uniqueness of the language faculty, Miller quotes an interesting metaphor from Samuel Butler in 1890 (from an essay "Thought and Language") which is worth replicating:

although growling and barking cannot be called very highly specialized language, still there is a sayer, a sayee and a convenanted symbol designedly designedly applied. Our own speech is vertrebrated and articulated by means of nouns, verbs, and the rules of grammar. A dog's speech is invertebrate, but I do not see how it is possible to deny that it possesses all the essential elements of language. p.87

Since Communication Theory is allied to Artificial Intelligence, there is a brief discussion of the achievements of computing and its origins in Turing's work in formal logic. I was impressed with how Miller keeps the tone of the book light without vulgarizing the subtle shades of his work and others. Here the much-misunderstood hypothesis of "thinking machines" is recounted correctly:

Several years ago the Englsh mathematician A.M.Turing considered the difficult question of whether or not a computing machine can think. Since the semantic and metaphysic issues are apparently unresolvable, Turing rephrased it. Can a computing machine, he asked, behave in the way we behave when we say we are thinking? This rephrased question, he felt, might have an answer, and to make the issue perfectly definite, he proposed what he called the "imitation game." [...] And that [experiment], he implied, is all anyone should ever mean by the question "Can machines think?" Turing, writing in 1950, predicted that within fifty years it would be posssible to build and programme computers that could do well at the imitation game. p.96-7

In this same section on the future of computing machines, Miller makes a comment which makes eerie reading in 2014:

What new horizons will this open up? Predictions are always hazardous, but I believe it means that the next generation of machines will be modelled along hte lines of what in the United States we call a public utility. Just as we can plug our electrical appliances into the wall and draw power from a central station, so in the future we shall be able to plug our typewriters into the telephone system and draw intelligence from a central computer. p.123

The last essay is the longest, and is a self-deprecating history of another research project called "Project Grammarama." Co-conceived with Noam Chomsky ("logician and linguist," as Miller honors him), the project tested the ability of average human beings to puzzle out the rule-governed systems of context-free grammars, that is, strings of symbols which are "well formed" or not according to principles such as X can be replaced by YZ. If X is replaced with ZZ or YY or &% then the string is not well formed. Given the chance to write their own strings, participants in the experiments had to recognize the rules only from being told that their strings were "CORRECT" or "INCORRECT." Miller is honest about the high expectations that he had for the project at the start, and admits that by these intial standards the project was a failure:

When we first conceived of Project Grammarama there was, I must confess, some thought in my mind that these experiments would cast light on the way infants acquire the rules of their native languages. It did not take long to disabuse me of this idea. There are many differences; taken together they add up to the conclusion that there is almost nothing in common between the task we set our subjects and an infant's problem of learning to talk. For one thing (1) our subjects were not infants. They were young adults who (2) already knew one or more languages; a second language is never learned the same way as a first. (3) The artificial language was visual, not auditory, which accentuates a different kind of patterning. (4) Our experiments did not involve meaning and (5) there was no use for the language once it was learned. (6) Infants work steadily for two years or more at their task, whereas our subjects had only an hour to two to give us. (7) The infant's goal is, of course, much more complicated than our subjects', but then (8) he seems to have an innate predisposition for solving it that was not apparent in the work of our subjects. (9) The infant, of course, is acquiring a sensori-motor skill, whereas our adult subjects were puzzling over an abstract cognitive pattern. In retrospect, the notion that our machine might play "mother" to some twenty-year-old "infants" seems embarrassingly naive. [...] It would please me if some practical application developed. Still, I have not been willing to push for it. In general, I think the wiser course is to minimize any attempts to transfer psychological principles between the learning of natural and the learning of artificial languages. p,173-4

Though I admire the clear thinking (as also shown in his liberality with the "communications" psychology metaphor), this last comment seems overly safe to me. Even if the observable process of a grown adult puzzling out artificial languages in the "Chomsky Hierarchy" (something I still don't understand) is not "tranferable" to the process of a child organically acquiring her first language over many years... is there not an apparent connection that seems to scream for comparison?

Finally, Miller says exactly one thing that I violently disagree with in this book:

My own interest in [Project Grammarama] was to discover how people acquire and obey systems of rules. By and large, they do not learn syntactic rules by memorizing particular examples of grammatical strings. To put the difficult in familiar terms; there may be some virtue to memorizing German sentences as part of your programme for learning the German language, but if that is all you do, you will surely learn a very odd kind of German. If people do not ordinarily try to learn linguistic rules by memorizing particular examples that conform to them, how much of their time should psycholinguistics spend studying the process? p.138

This is simply incorrect. It does not improve your German to learn the rules first more than it does to learn after memorization of "particular examples" (search Alexander Arguelles's "Shadowing" method of language learning). The leap of imagination that Miller lacks is of the process by which memorized examples of grammatical examples become recombineable in your head. If his experience is of learning grammar, exercising in grammar and then finally noticing examples, this may explain this narrow-minded viewpoint on language learning, but also obviously it stemms from the way natural language is first learned. I appreciate that Miller is only saying here that some principles for recombining linguistic building blocks surely come before the examples... but I think that someone with experience in foreign languages would avoid falsely drawing equivalents between L1 and L2 acquistion. The first is an organic part of cognitive maturation and the second is some diligent training in the capacity of analogy and "chunking" informational units.