How Did the “Street Graphics” Book Misrepresent Dr. Miller’s Research?

A subject being irresponsibly employed to restrict sign size at the local level

By Dr. R. James Claus

(This article was originally published in the April 1973 issue of Signs of the Times magazine. )

Following are quotes from a letter from Dr. George A. Miller, the well-known experimental psychologist. The letter is largely self-explanatory. Dr. Miller refers to Chapter 3 of the book Street Graphics by William Ewald and Daniel Mandelker, wherein the authors are discussing information overload and have cited some of Dr. Miller’s previous work in support of their points.

Dr. Miller’s letter also treats some of the issues underlying the study of human information-processing and refers to the misinterpretation of such research by persons who inadequately understand the psychological processes involved in the phenomena they are seeking to regulate.

One of the popular arguments advanced by those who seek stringent regulation of signage is that signs are “overloading the information system” and that this overload constitutes a hazard to general public safety, health and welfare. They argue that the human observer cannot process the information presented on signs unless copy area and content are severely limited. Therefore, it is argued, the conscientious planner should limit the number of “items” (defined in convenient terms) on a sign in order to lessen the confusionand frustration of the average American citizen which is caused by the overload of information presented by the commercial community. To sustain such control, Dr. Miller’s classic 1956 article is cited (or miscited as you will see) as the scientific evidence supporting the rationale.

In fact, as you will note by comparing the article with what is said in Street Graphics, Dr. Miller clearly indicates that we know little about how a human processes information, and more important, this processing is greatly dependent upon the background and experiences of each individual observer. What is a “chunk” of information for one person may break into a number of “bits” for another. The important point is that the experts in the field of informationprocessing would agree on one primary fact: we know too little about the phenomena in question to even attempt regulation.

I sincerely hope that as the members of the sign industry begin to see planners produce reports referring to “information overload,” the enclosed letter will be brought to the attention of the persons involved.

With respect to most matters discussed in the book, (Street Graphics), I am not competent to comment. Chapter 3, however, on “The Physiology and Psychology of Seeing,” treats a subject I understand, and some of my own work is summarized on page 20. Since that summary misrepresents what I wrote in 1956, and since legislation based on that summary could have important implications for efforts to improve street graphics, I would like to try to clarify the passage in question.

Since the precise wording involved is important, I will begin by quoting the three sentences I object to most strongly:

Regarding the human being’s fundamental limits in processing information, Princeton Psychologist George A. Miller has concluded from various experiments that we possess a finite and rather small capacity for making judgments about the information we receive through the operation of our senses, and that this capacity does not vary a great deal from one sense to another (hearing, sight, etc.) Miller defines this capacity as the span of absolute judgment, and his findings indicate that the limit of this span is about seven categories of information. This is to say that the average observer cannot distinguish between more than seven different sights or sounds presented to him simultaneously.”

In order to illustrate the misrepresentation that is involved, and to protect my own reputation as a scientist, I must quote more fully from that paper.

First, however, let me say that, although I live in Princeton, N.J., I have never been a member of the faculty of Princeton University. The rest of the first sentence seems to have been taken from my paper, but with one important omission. The original paper said:

“On the basis of the present evidence, it seems safe to say that we possess a finite and rather small capacity for making such unidimensional judgments and that this capacity does not vary a great deal from one simple sensory attribute to another.”

In Street Graphics the adjective “unidimensional” has been omitted, and my phrase, “from one simple sensory attribute to another,” has been changed to read, “from one sense to another.” At first glance, these might seem to be minor editorial changes to make the writing more intelligible to a lay reader. But the modification completely changes the meaning. Indeed, it changes the sentence from one that I still believe might be true into one that is certainly false.

As my paper makes clear enough to a careful reader, the word “unidimensional” means that the stimuli the observer was asked to identify differed from one another with respect to only one sensory attribute. A sensory attribute, which the author of Street Graphics has confused with a sensory modality, is a characteristic of a sensory experience. The attributes generally recognized are quality, intensity, duration, and extensity; for visual sensations, for example, these attributes are color, brightness, duration, and size. Perhaps my terminology was too technical for non-psychologists, but I still know no better way to state the true situation than to quote my next paragraph, following the sentence quoted above:

“You may notice that I have been careful to say that this magical number seven applies to one-dimensional j u d g m e n t s. Everyday experience teaches us that we can identify accurately any one of several hundred faces, any one of several thousand words, any one of several thousand objects, etc. The story certainly would not be complete if we stopped at this point. We must have some understanding of why the one-dimensional variables we judge in the laboratory give results so far out of line with what we do constantly in our behavior outside the laboratory. A possible explanation lies in the number of independently variable attributes of the stimuli that are judged. Objects, faces, words, and the like differ from one another in many ways, whereas the simple stimuli we have considered thus far differ from one another in only one respect.” The second sentence in the passage I have quoted from Street Graphics seems to be a paraphrase of the following statement in my 1956 paper:

“There is a clear and definite limit to the accuracy with which we can identify absolutely the magnitude of a unidimensional stimulus variable. I would propose to call this limit the span of absolute judgment, and I maintain that for unidimensional judgments this span is usually somewhere in the neighborhood of seven.”

Here, again, Street Graphics has omitted “unidimensional” in what they have borrowed from me. Lest you think I am quibbling about style, let me say again that the omission of “unidimensional” transforms what I actually wrote into the absurd claim that there are only seven different things in the whole world that anybody can identify perceptually. No doubt I have said some foolish things in my life, but this is not one of them.

The third sentence in the passage from Street Graphics is hard to understand and I can find nothing in my paper from which it could have been taken. The experimental situation it seems to describe does not correspond to the experiments on absolute judgment that my paper discussed. In an experiment on absolute judgment, one stimulus is presented at a time; the observer is asked to make one out of a variety of possible responses in order to identify for the experimenter which stimulus he thinks occurred. Absolute judgment contrasts with relative judgment. In an experiment on relative judgment, two stimuli are presented at a time; the observer is asked to judge one relative to the other as brighter, redder, larger, louder, etc. Concerning methodology, the 1956 paper said:

“There are two ways we might increase the amount of input information. We could increase the rate at which we give information to the observer, so that the amount of information per unit time would increase. Or we could ignore the time variable completely and increase the amount of input information by increasing the number of alternative stimuli. In the absolute judgment experiment, we are interested in the second alternative. We give the observer as much time as he wants to make his response; we simply increase the number of alternative stimuli among which he must discriminate and look to see where confusions begin to occur.”

Perhaps a reader unfamiliar with this kind of research might think that “increasing the number of alternative stimuli” meant increasing the number of “different sights or sounds presented to him simultaneously.” This interpretation would be incorrect. The one particular stimulus that did occur on any given trial was chosen from an increasing number of alternative stimuli that might have occurred, but didn’t.

It is an interesting feature of perceptual recognition that accuracy depends not only on the physical characteristics of the object to be recognized, but also on the size of the set of alternative objects that the observer believes might have occurred instead. For example, the word “dog” may be heard correctly at very faint intensities when the listener knows in advance that the word he will hear must be either “dog” or “cat.” But if all he knows in advance is that the word he will hear must be an English monosyllable, the same word “dog” must be about twenty times as intense in order for the listener to hear it correctly. In this and several other publications during the 1950s I pointed out that this feature of perception could be explained in terms of the engineer’s probabilistic definition of “amount of information.” The number of “bits of information” that an observer can extract from a degraded stimulus-a word spoken over a noisy telephone, or a word printed in low contrast and briefly viewed–seems to be roughly constant and determined by the degree of degradation. When the identification he is asked to make is easy, however, a few bits of information may suffice, but when he doesn’t know what to expect, when “anything can happen,” the information he can extract from a degraded stimulus will not be sufficient to support accurate recognition.

In short, it is impossible to predict from the physical properties of a perceptual situation what an observer will be able to identify correctly. Such predictions must take into account the observer’s uncertainty about what he expects to see.

Such was the real message in the first half of my 1956 paper. In the second half I went on to argue that the short-term memory differs from perception in just this respect. When we measure how much of a message an observer can remember after hearing or seeing it only once, we find that the number of alternative messages he might have been asked to remember instead is irrelevant. All that matters is the length of the message. Why the perceptual system should be characterized by one set of invariants and the memory system by another was the puzzle I wished to pose for my colleagues in 1956. Subsequent research on short-term memory has demonstrated that it is far more complicated than could be revealed by simple experiments on immediate memory, but that is another story.

My 1965 paper does argue, however, that memory is facilitated if the message to be remembered can be grouped into familiar units, which at the time I called “chunks.” Whereas “bits of information” is a well-defined concept, “chunks of information” is far more subjective and ill-defined. The point is this: if memory depends on length of message, how should length of message be calculated? For example, should “Drink Coca-Cola” be counted as 13 letters long, three words long, two words long, five syllables long, or one phrase long? The answer cannot be determined from the formal properties of “Drink Coca-Cola,” but must take into account the rememberer’s prior experience with such messages.

Someone who knew the alphabet but no words of English would have to treat it as if it contained 13 chunks of information, in which case it would exceed his span of immediate memory. Someone who knew English but had been shielded from advertising might treat it as if it contained three chunks of information. Those of us who have seen this slogan many, many times could treat it as a single chunk. Exactly the same physical stimulus will have different numbers of chunks for different people.

These considerations may seem academic and unnecessarily technical, but I believe they are relevant to another passage on page 32 of Street Graphics where the author discusses “Information Load”:

“In the example given for calculating total graphic area, the number of letters was assumed to be 30 because this is approximately the flu m be r — needed for a street graphic system that displays ten items of information — i.e., ten syllables, symbols, abbreviations, broken pianes, or discontinuous odd shapes. A major factor in street graphic overload is that frequently the viewer is confronted with more items than his eye can see or his mind can process (see Chapter 3). The number of items that can be viewed at one time from a moving car is a function of driving speed, traffic conditions, and graphic design. Results of the Baltimore County study indicate that, under normat driving conditions, the occupant of a moving car can seldom handle more than ten items of information at a time. From this limitation proceed the rules given in the following chapter [or determining how much information an individual premise may display.”

I find it ironic that my 1956 paper should be cited to support such recommendations. The whole point of that paper was that “items” of information were insufficient. It is not the number of items, but the iiumber of bits of information that will determine what he can see, and the number of bits depends on his expectations. It is not the number of items, but the number of chunks of information that will determine what he can remember, and the number of chunks depends on his previous experiences. The situation would be amusing if the misrepresentations of my work were not being taken as the basis for enacting ordinances to control street graphics.

In general terms, I approve of the goals of Street Graphics. But I must strongly protest the distortion of my own work, and must deplore the enactment of restrictions based on such an inadequate understanding of the psychological processes involved.

I hope these comments will help to clarify the relevance — or, more accurately, the irrelevance — of my 1956 paper for the fascinating problems treated in Street Graphics.

George A. Miller, Professor — The Rockefeller University

 

Photo credit: msbritt

Posted in Blog: Rhetorical, Research, Universities, Visibility and Legibility.