Ye Olde Natural Philosophy Discussion Group

Reviews and comments on
Douglas Hofstadter & Emmanuel Sander:
Surfaces and Essences: Analogies as the Fuel and Fire of Thinking [2013]

This book attempts to prove, primarily through a huge mass of examples of quite varied kinds, with regard to words, phrases, metaphors, proverbs, “me, too” stories, objective situations, and so forth, that the main way that people think is through analogies. It also argues that another way of saying more or less the same thing is that categorization is the grouping of things and ideas on the basis of analogies.

One person in our group (Scott) thought that this was an important and useful book. His review appears below. Kirby also liked it and read the entire large volume. He commented that Hofstadter is known for repeating himself. And so Kirby did sometimes just skim the extra examples that the authors adduced to prove their points. Kirby had already read a book like this one, which at times presented what seemed like a half-baked theory, so this book expanded upon that one. He felt this book puts forward a great theory, but critized it because he just didn’t see the great importance of what they were saying. “What makes me care about the point [of the book]?” Still, he rated the book a 7 on our scale of 0 to 10.

However, just about everyone else in our group was quite negative about the book, didn’t read very much of it at all, and gave it very low ratings. Our average rating was just 2.8.

Ron rated it 0. He said he had a real problem getting into it, and didn’t like the examples. He eventually got to the point of only reading the first sentence of every paragraph because of the huge number of examples.

Kevin gave it a 1. He tried listening to an audio version of the book. “If reading it was as bad as listening, I’m sorry. I liked the concept, but the writing wasn’t accessible. It seemed like a textbook, a terrible one.”

John also gave it a 1. He started reading it, got through the prologue and then chapter 1 felt like more of the same content. He felt like we had read other books that covered the concept. The authors just kept repeating the word analogy over and over. “I couldn’t gain traction and gave up after chapter 1.”

Rich, too, gave it just a 1. He only read 75 pages and then gave up. He called it “one of the most dense, repetitive books we’ve read”, and agreed with many of the other negative comments about the book.

Vicki rated it a 2. She did not finish the book, and stopped about halfway through. “I liked the concept. I think the theory is solid. I really enjoyed the prologue and first chapter. However, the writing is dense, and it becomes increasingly tedious. The examples beat each point to death (to use an analogy...). But what really killed it for me, was the implicit bias shown in the examples. The assumption that traditional gender roles are universally understood, the use of offensive stereotypical analogies, such as the Jewish Mother, and the ego of the two authors portrayed in some sections, all contributed to ruining any potential for me to enjoy it. I gave it 2 stars for the concept, but this book needs a serious editor that can cut it down and remove the biases.”

Barbara rated the book a 5. She read the whole book, but said a lot of it is garbage. “What do the examples have to do with science?” she asked. “A lot of sayings are in the book, but it’s too much and it needs editing.”

Rosie didn’t have a way to get to the meeting, and missed it. However, she sent out a short email about it which had similar negative views of the book, and also rated it a flat zero.

Scott missed the meeting but here is his review:

I brought this book to the attention of our science book club because, while I had already long ago come to believe that analogies are of central importance in how humans think, I recognized that for me this still remained somewhat of an undeveloped thesis (though not just some kind of wild hunch or guess). Moreover this belief was probably mostly based on introspection and on occasional isolated comments by philosophers and artificial intelligence researchers, such as the following remark by Marvin Minsky:

“How do we ever understand anything? Almost always, I think, by using one or another kind of analogy—that is, by representing each new thing as though it resembles something we already know. Whenever a new thing’s internal workings are too strange or complicated to deal with directly, we represent whatever parts of it we can in terms of more familiar signs. This way, we make each novelty seem similar to some more ordinary thing.” [Society of Mind (1986), p. 57.]

Yes, that does seem like a very plausible claim. And, as I say, I was convinced of the idea at least since the time I read that passage. But I still vaguely felt that my adherence to this theory was based on insufficient investigation. I wanted to get a lot deeper into the matter, hear some serious arguments for the idea, and see a lot more evidence—as well as consider any criticism of the claim that analogies are the basic way of thinking. Did this book do the job? Well, yes and no! On balance I give the book a rating of 8.

The basic method used by the book is to bury the reader in an enormous number of diverse examples of analogies used in ordinary conversation and activity, in order to prove the ubiquity of analogical thinking. That sort of makes sense. If the claim is that analogy-making is everywhere, then to prove it you sort of have to look everywhere. But often, when going through the extremely long lists of examples of some particular kind or use of analogy I did start to think to myself, “OK, I take your point. Let’s move on...” But then I remembered that I already agreed with their overall thesis before starting the book; maybe all this mass of proof really was a good idea for those who were not very sympathetic to the centrality of analogy in human thinking before they even opened the book’s covers. The amount of proof of some thesis which has to be provided does in fact depend upon the reader’s background and presumptions.

However, I do suspect that many readers will still find the book as way too long. It is no doubt true that the theoretical points in the book could be summed up in just a couple dozen pages at most. But would the authors then have made the same powerful case they do? No, they wouldn’t have. It is the great mass of evidence that mostly constitutes their case.

The second most central theme of the book (after the claim that analogies are the basic mode of human thought), is that categorization is essentially the same thing; i.e., that we ascribe specific things to definite categories based on their analogies with other things already deemed to be included in those categories. The authors seemed to feel that they had had difficulties of proving their point on this issue, and returned to it several times, including in the “Epidialogue”. In particular, they repeatedly argued for the somewhat unclear assertion that categorization is not a mere dropping of a new thing into an existing categorical box, but “rather” the making of an analogy. (Sometimes I sort of understood them, and other times I wondered why both things could not be the case!)

A “category” is equated with a “concept”, at least for the most part. Most people would probably feel more comfortable with the word “concept”, but the authors were looking for a more general term that was not exclusively psychological. It is interesting (and very likely unknown to the bourgeois authors) that Marxist philosophical theory did the same thing long before. (It is true, though, that the Marxist notion of categories focuses on what the authors would consider only the relatively most prominent categories—the major categories of thought and reality—and generally ignores the plethora of trivial cases.)

Speaking of politics, many of the examples offered by the authors of the various analogical situations are often slued toward their academic upper-middle class milieu. But unlike many of the “science” books our group has read, there are refreshingly few gratuitous slams against communism and revolutionary figures.

I thought the more abstract discussions of analogies related to the “Copycat Domain” were one of the weaker parts of the book as far as readibility goes; too “mathematical” I guess. But overall the book was easy to read, and often a real delight. I read many of the short examples and passages to my wife Sara as I went along. I thought the last chapter, which focused on the analogies that Einstein used to develop his revolutionary theories in physics was fascinating, as was the earlier description of how analogies between water waves, light waves, and other types of waves developed analogically in science.

But on the more negative side I hoped that there might be some description, or at least hypotheses, put forward about how it is that the human brain can so easily make analogies. Are there one or more “analogy engines”? These days it seems that this line of investigation would naturally lead into how neural networks work, AI, “deep learning”, and so forth, especially in connection with the authors’ discussion of topics like “conceptual encoding” (p. 172) and “brainbows” (pp. 182-3). Maybe this further topic was considered too big of a subject to include in the book, and deserving of its own book (or books). Still it surprised me that virtually nothing was mentioned about all this here.

There are some conceptual areas which very much seem to be deep and profound, but where the discussion of them somehow never seems to be complete or get to their deepest essence. One example of this is the topic of emergence in science, where new phenomena emerge out of simpler phenomena. (Such as temperature and pressure “emerging” out of the massive flux of simple moving molecules.) Often the very fact of existence of emergent properties is not understood at all. Arthur Conan Doyle has Sherlock Holmes say: “From a drop of water a logician could infer the possibility of an Atlantic or a Niagara without having seen or heard of one or the other. So all life is a great chain, the nature of which is known whenever we are shown a single link of it.” (From A Study in Scarlet.) This is quite incorrect. Waves, waterfalls and oceans are not “logical implications” of a drop of water (at least as far as an ordinary human’s acquaintance with such a drop goes). On the contrary, these are emergent phenomena couched in their own categories and governed by their own emergent dialectical contradictions.

So I am making an analogy here: It seems to me that discussions of analogies themselves and of analogical thinking tend to be like discussions of the topic of emergence: namely something still a little unsatisfactory, still not fully explicated, still not complete, still a little disappointing. As good and useful as Hofstadter & Sander’s book is, it still seems to me that there is more to be said about analogies and their nature and aspects. I thirst for more, perhaps in a somewhat different vein.

On the other hand, I’ve had feelings like this on various topics before which eventually proved unfounded, as my brain got reorganized about some subject. If I’m still around a few years from now you can check to see if whatever cognitive dissonance there may be in my thinking about analogies has dissipated!   —Scott

P.S. Because I thought there were numerous well-expressed or profound comments in the Hofstadter/Sander book I added a number of quotes from it to my “Dictionary”, including:

Return to our complete list of books.

Return to our Science Group home page.