On Misrepresenting Lloyd Morgan's Canon:
Comment on Thomas

Donald A. Dewsbury
University of Florida

dewsbury@ufl.edu


I appreciate the efforts of Roger Thomas (2001) in publicizing what I and other authors have been writing for a number of years: that is that Morgan's canon has often been mis-represented. That said, I disagree with Dr. Thomas about some aspects of this issue, particularly in relation to his representations of my writings. It appears as though we must agree to disagree.

I would like to make just a few points; I have no wish to get into an extended debate, which I think would be counter-productive. Dr. Thomas and I are basically on the same side, lamenting the mis-representation of Morgan's canon. We quibble over a couple of words.

As Dr. Thomas points out, during my mis-spent youth, I accepted and repeated what one of my teachers, P. L. Harriman, and others had taught me was the meaning of Morgan's canon. In essence, they contended that Morgan's canon and the law of parsimony, roughly equivalent to Occam's razor, were one and the same. Somewhere around the late 1970s I actually read Morgan's book and Newbury's (1954) article and discovered that the canon and the law of parsimony are different. Ever since, I have tried to correct the view that they are not.

Thus, in my 1984 book, I wrote that the prevailing view was incorrect:

"Morgan's canon refers to the level of psychical faculty or mental development assumed when explaining behavior. It refers particularly to comparisons across species and was related to the theory of emergent evolution. Occam's razor and the law of parsimony relate not to psychical levels but to the number of assumptions used in explanation" (pp. 187-188).

In a more recent article, faced with severe constraints on space, I wrote that:

"In his classic textbook, Morgan (1894) outlined his famous canon that an animal's behavior should be interpreted in terms of the psychologically simplest processes consistent with the data. Morgan's canon, and its related concept, parsimony, spread widely during this period" (p. 751).

Dr. Thomas cites both works as examples of distortion of Morgan's canon. I agree with most of what Roger Thomas, in the tradition of Edward Newbury, Rob Wozniak, and others has written elsewhere in his article. However, we must agree to disagree on two points concerning his presentation of my views. Basically, I think we are really quibbling over the use of a couple of words.

Point 1: Simplicity.

How should the concept of "simplicity" be used? Dr. Thomas objects to any use of the notion of simplicity, or what he calls "the simplicity criterion." I believe that the applicability of adjectives "simple" and "complex" depends on the noun which is being modified by these adjectives. I distinguish between simple vs. complex explanations and simple vs. complex psychological processes. The former is relevant to parsimony; the latter, I believe, refers to Morgan's canon. Complex explanations employ many assumptions and elements. As Dr. Thomas points out, Morgan is vague in writing of "higher" and "lower" faculties. Surely, he is using "higher" and "lower" in a metaphorical sense. Metaphors referring to either height or complexity are still metaphors and imprecise. In my view, both are imperfect. Dr. Thomas may have a point in that "simplicity" may be no better than "higher" or "lower;" however, I think it no worse. I tend to use the two almost interchangeably. I believe this consistent with Morgan's usage. It seems to me quite reasonable in trying to put Morgan's notion into words that might be meaningful today to apply the terms "simple" and "complex." The important distinction is that the adjective refers to simplicity OF PSYCHOLOGICAL PROCESS and not of simplicity OF EXPLANATION. Again and again, Thomas criticizes the use of simplicity of explanation as exemplifying Morgan's canon; I agree with him.

I agree with Dr. Thomas that Morgan is vague in what he means by "psychological scale." At a crude level, I think that most people would agree that there is a difference between a taxis and rational thought; there is some meaningful, if vaguely defined, dimension there. It may have been fine to use "psychological scale" to describe this dimension in Morgan's time. I don't think it is today. Most comparative psychologists and zoologists are trying to get colleagues and students not to think in terms of single scales, as they suggest an unrealistic conception of evolution that stems from Aristotle rather than Darwin. Thus, unless constrained by serious space limitations, I generally, where space permits, try to quote Morgan verbatim and then interpret him in words better suited to our time. I think this is necessary to get Morgan's point across in our context. Although I agree with Dr. Thomas that simple vs. complex is not an ideal dimension with which to convey Morgan's meaning, I cannot think of a better one. Perhaps one could use "evolutionarily older" vs. "newer" but such usage has excess baggage too. If we can find other words with which to express this, I would welcome them. Until then, I plan to continue writing about psychological complexity as the best term I can find.

My concern stems, in part, from the Scala naturae issue. I am concerned that the notion of higher and lower processes will be confused with the notion of higher and lower species. We have worked hard to get notions such as "higher versus lower" species, "sub-humans," etc. out of the vocabulary of psychologists. As many believe different levels of psychical processes are characteristic of different species, there is a risk of a return of such outworn concepts. As we no longer accept notions of a single great chain of being, I favor an alternative wording for Morgan's idea if it is to be more useful today.

Elsewhere in his 1894 book, Morgan also used the terms "simple" and "complex" in the sense that I do. In a telling sentence that Dr. Thomas actually includes in his long quotation, Morgan wrote:

"Moreover, the simplicity of the explanation of the phenomena of animal activity as the result of intellectual processes, can only be adopted on the assumption of a correlative complexity in the mental nature of the animal as agent" (p. 55).

Here, Morgan is showing us exactly the distinction I wish to draw between complexity of explanation and complexity of process or "mental nature of the animal." This passage appears shortly after Morgan's declaration of the canon in a section where Morgan is differentiating the canon from parsimony. He distinguishes between "simplicity (or complexity) of the explanation" and "complexity in the mental nature of the animal," or, as I would say, of the psychological processes of which an animal is capable. I read Morgan as saying rather clearly that parsimony is not an adequate criterion for assessing the adequacy of an explanation unless it is correlated with psychological complexity of the animal in question. We should not accept the simplest EXPLANATION consistent with our observations but the explanation that relies on postulation of the simplest PSYCHOLOGICAL PROCESSES.

Morgan applied the concept of complexity to psychological processes in a manner similar to my use elsewhere in the book as well. Earlier (p. 8) Morgan wrote:

"Given an organism in which analysis gives two aspects, complex energy and complex consciousness, from what have these been evolved by an evolution which is selective, synthetic, and cosmic or determinate? From the nature of the case, the bodily aspect is that of which alone we can have objective knowledge. We trace the evolution backwards and find, in our interpretation thereof, simpler and simpler organisms, until the organic passes into the inorganic. We find the energy less and less complex as we look back through the vista of the past. And what about the other aspect? Does it not seem reasonable to suppose that, no matter what stage we select, analysis would still disclose the two aspects? That with the simpler modes of nerve-energy there would go simpler modes of consciousness, and that with infra-neural modes of energy there would be infra-consciousness, or that from which consciousness as we know it has arisen in the process of evolution?"

In this paragraph, Morgan used the adjectives "simple" and "complex" to modify the nouns "energy," "consciousness," "organisms," "nerve energy," and "modes of consciousness." Whereas Morgan uses the concepts of simple and complex in such ways, particularly where he refers to "simpler modes of consciousness," is it really unreasonable to apply it in relation the the canon?

Point 2: Relatedness.

Are Morgan's canon and the law of parsimony related or unrelated? I believe that the two are related. Both are concerned with the issue of the construction of adequate explanations of animal mind and behavior. One concerns psychological processes; the other concerns assumptions and is of broader application. In my view, these are related, though different, concepts. Dr. Thomas seems to believe that they are totally unrelated. Of course, whether two concepts are related or unrelated is a matter of degree that rests, in part, in the eye of the beholder.

Dr. Thomas objects to my wording in the article with severe space limitations where I wrote of "Morgan's canon, and its related concept parsimony" (Dewsbury, 2000, p. 751). I believe that the two are related. Indeed, if they are not, I don't understand why they are so often confused! Thus, I stand by my statements of the last 20 years or so that Thomas quoted, including brief statement of 2000 "In his classic textbook, Morgan (1894) outlined his famous canon that an animal's behavior should be interpreted in terms of the PSYCHOLOGICALLY SIMPLEST PROCESSES consistent with the data. Morgan's canon, AND ITS RELATED CONCEPT, PARSIMONY, spread widely during the period" (emphasis added), and my 1984 statement "the law of parsimony and Morgan's canon are two CLOSELY RELATED principles" (p. 188).

In order to be "related," two things must, by definition, be different. Thus, in my view, when we say that two things are related, we are, of necessity, saying that they are different.

Notes in Passing.

Let me note Wozniak's (1993) contribution in really emphasizing the role of introspections and consciousness in Morgan's thought. Although I touched on this ("the canon was not written in an effort to eliminate the attribution of consciousness to animals. Morgan assumed the existence of consciousness in animals" (Dewsbury, 1984, p. 188), I did not bring it out in the forceful manner of Wozniak.

Let me also note in passing for those interested in Morgan's canon an interesting article by Gregory Radick (2000) on the origins of the canon in relation to the work of R. L. Garner.

Conclusion.

In summary, Dr. Thomas and I agree about the misinterpretation of Morgan's canon; we disagree about the utility of the terms "simple vs. complex" and "related vs. unrelated." Further, to add a third word, I think it inappropriate for Dr. Thomas to list my wording as a "distortion" of Morgan's canon. There are aspects of my treatment of Morgan's canon in both my 1984 and 2000 discussions that, in the interests of clarity, I would change today if I could; neither is perfect. However, in my writings of the last 20 years or so and in Thomas's recent writings, we agree that Morgan's canon has been mis-represented. I believe that we are in a basic agreement as to what Morgan meant. We choose to use slightly different words in expressing our views. It would appear that we must agree to disagree. Perhaps, each in our own way, we both can work toward an accurate understanding of this important principle.


References.

Dewsbury, D. A. (1984). Comparative psychology in the twentieth century. Stroudsburg, PA: Hutchinson Ross.

Dewsbury, D. A. (2000). Issues in comparative psychology at the dawn of the 20th century. American Psychologist, 55, 750-753.

Morgan, C. L. (1894). An introduction to comparative psychology. London: Walter Scott.

Newbury, E. (1954). Current interpretation and significance of Lloyd Morgan's canon. Psychological bulletin, 51, 70-74.

Radick, G. (2000). Morgan's canon, Garner's phonograph, and the evolutionary origins of language and reason. British Journal for the History of Science, 33, 3-23.

Thomas, R. (2001). Lloyd Morgan's canon: A history of misrepresentation. History & Theory of Psychology Eprint Archive. Article htp00000017. Retrieved from http://htpprints.yorku.ca/documents/docs/00/00/00/17/htp00000017-00/MCWeb.htm.

Wozniak, R. H. (1993). Conwy Lloyd Morgan, mental evolution, and the Introduction to Comparative Psychology. In Morgan, C. L. An introduction to comparative psychology (reprint edition 1894/1993) pp. vii-xix). London: Routledge/Thommes.