Surfaces and Essences: Analogy as the Fuel and Fire of Thinking by Douglas R. Hofstadter

Surfaces and Essences: Analogy as the Fuel and Fire of Thinking Reviews

• 
  Ruben

  Reading
  Surfaces and Essences: Analogy as the Fuel and Fire of Thinking is like biting an apple and finding half a worm. It's like a pancake eating contest. It is the Phantom Menace of cognitive science literature. What should have been a monumental work about understanding via analogy undermines itself by being too repetitive, too unfocused, too obvious, too silly and too self-referential.
  
  I wanted this book the minute I saw the title because I'm a big fan of well-crafted analogies. I remembered hearing good things about Dr. Hofstadter's book,
  Gödel, Escher, Bach: An Eternal Golden Braid from college roommates who'd read it, which added to my sense of anticipation.
  
  Sadly, other than the prologue and parts of the final chapter, I find very little to recommend here.
  
  The book opens with an exploration of the "zeugma", which is the use of a single word in two different ways in the same sentence. An example of a zeugma from a song I wrote is "I can make you a cup of tea/And you can make me smile.” This begins to get at the ability of the human mind to make and break lexical categories in unexpected ways.
  
  Yet starting with the first chapter, Dr. Hofstadter and his co-author, Emmanuel Sander, seem intent on removing everything that was interesting about analogies by taking a particular word or expression and overanalyzing its figurative meanings for a number of pages. The reason this comes across as extremely tedious is that the point has already been made, and it’s easily understood. No one who knows what an analogy is needs to ruminate over why a mother board is a little bit like a real mother. Most people who read this book will go dozens of pages at a time without learning anything new.
  
  A number of pages compare airport "hubs" (for airlines) to the hubs of wheels. Who cares?
  
  There's a section on the difference between "and" and "but", as if anyone on Earth could have reached page 109 without understanding that.
  
  They bother to point out that understanding is not actually standing under anything.
  
  One particularly unreadable section uses letter patterns to illustrate how we are reminded of past events. They somehow thought that people wanted to read about how iijjkk-->iijjkd is different from iijjkk-->iijjd and iijjkk-->iijjll. This reminded me of how in high school a classmate thought that the page numbers had some connection with what was happening in the novel. I can't bring myself to care about that.
  
  Chapter 7: Naïve Analogies and Chapter 8: Analogies that Shook the World were the only two worth writing, though they were not particularly well done. Naïve Analogies discusses the ways that using physical terms to explain abstract concepts limits our understanding. The best example is that people think of division as splitting something up, thereby making it smaller. However, this is a limiting analogy. If you have four bags of chocolate chips, and you need half of a bag to make one batch of chocolate chip cookies, you can make eight batches of chocolate chip cookies, ending up with a number (8) that is larger than the first two (4 and .5). Thus, 4 ÷ .5 = 8. The last chapter explored analogies in physics, concentrating primarily on Einstein’s theories. Unfortunately, the language was so abruptly technical that it seemed like it had been written by a different author.
  
  The Epidialogue has to be the worst possible way to end a book. It’s a made-up conversation about two friends discussing categories vs. analogies and referring to parts of this book, Surfaces and Essences, including this very epidialogue. Then one of the characters wakes up, and has a conversation about the crazy conversation it had just dreamed. Then one of the characters wakes up, and it turns out that the second conversation was all a dream, too. It’s bad.
  
  Lastly, (analogy alert) the authors’ rampant use of clunky, mixed metaphors reminded me of Stephen King’s “dandelions” from
  On Writing. Dandelions are unobtrusive until you notice them, and then you really notice them, and they get on your nerves. (For Stephen King, dandelions are adverbs, as in “she replied nonchalantly”). For example, Hofstadter and Sander write: “Once he had glimpsed this analogy, Einstein went way out on a limb, placing all of his chips on it, in a move that to his colleagues seemed crazy.” And “The turning point when light quanta at last emerged from the shadows came only in 1923.” I’m not sure if the authors are trying to be cute or if they just don’t proofread, but this stuff is not good.
  
  It makes me annoyed and write a long negative review.
• 
  Bradley

  There are two or three different books in this book, but by all apparent surfaces, it is all a single, exhaustive tome on ANALOGY.
  
  As I read it, I was struck by how vast and careful his analysis was and how I would have REALLY loved this as a teen, being fascinated by all the variances, categories, and richness of analogies. They are a source of amusement, creativity, vast and widespread accidents, a mode and end of consciousness, and an integral aspect of math and science. What is an equal sign but an analogy? And let's not forget Einstein making thought experiments that later became provable.
  
  Analogy is in every word we use, constraining and freeing our understanding of the world as well as tumbling it into a mass of contradictions. Only logic and careful analysis can free it, but the source of all our greatest creativity comes from it.
  
  As a kid, and perhaps unused to all the varieties of analogy and hungry for such a careful and well-thought-out stream of reasoning, I probably would have given this a full 5 stars just for is sheer chutzpah.
  
  As an adult, I think it went on WAY TOO LONG.
  
  Once the great and rather obvious arguments had been made, fought over, and survived the logic grinder, I was perfectly happy to throw it on the grill and garnish my the buns of my life with thrilling mustard, spicy onions, wholesome lettuce, and timely tomatoes.
  
  I could easily see this book fueling the understanding of our cognition or developing Artificial Intelligence. I can see it becoming a monumental if a rather pedantic tribute to obviousness. But it is obvious only because we're in the heart of it.
  
  Or rather, I might recommend this book to aliens trying to understand us humans. Or AIs from other stars trying to get a good grip on our alien psychology.
  
  For the general lay-reader, DESPITE it being always lightly humorous and clear, I cannot recommend this... except, perhaps, in small doses while sitting on the toilet.
  
  But am I happy I read it? Yes. I can say I'm a complete convert to the line of logic. It aligns to my own reasoning very well. But was it often boring AF? Yes. It was that, too.
• 
  Brian

  I read GEB on publication and it remains a landmark book for me. "The Mind's I", co-edited with Dennett, is fantastic and I greatly enjoyed his book on the perils of translation, "Le Ton Beau de Marot". Others might be hit and miss but always provided *something*. This one, co-written with Emmanuel Sander is not only an immense disappointment but just bafflingly bad. I'd read some reviews on Amazon that referred to certain problems but thought they might be exaggerating. What you have, at heart, is something that might have made a decent magazine article--might have--fleshed out to absurd lengths. I had the strong impression that virtually every paragraph was gone over with the idea: How can we bloat this thing? Why give one or two examples when forty will do? How about two or three entire pages of examples where everyone got the point within one line? And this happens *constantly*. It becomes an impossible to shrug off chore as a reader, to look for the padding--and you're "rewarded" almost every time. Add to this that the essential point is that thinking rests on the ability of the brain to do all sorts of analogies--not such a stirring thesis, imho--and you have a dismal, feeble excuse for a book. Shockingly bad.
• 
  Rossdavidh

  Douglas Hofstadter (writing here with his Francophone colleague Emmanuel Sander) has written a 5-star book of about 200 pages, and has hidden it inside a 500+ page tome. He has made a really splendid book, and then repeated each and every point so often as to make the whole thing longer, and less splendid. He has penned a truly excellent work, were it of a modest length, but he has engaged in an abundance of redundancy, such that the whole is less than the sum of its parts, were there fewer parts. He wrote a really fine work, that was greatly in need of a stern ruthless editor, not because any part of it was bad but rather because often four pages were used when one would do. He has...
  
  I think you get the point.
  
  To be honest, I gave up on this book about 1/3 of the way through, the first time. Then it sat on my "to be read" shelf for a year or two, waiting on judgement, in a sort of libro limbo. I had very nearly decided to send it to the used bookstore, but I did not, and ultimately returned to it.
  
  The reason is that, over time, I noticed that my mind returned to the ideas in it again and again, months or years after reading only the first part of it.
  
  So, once more unto the breach, and I started from the beginning because it had been so long. In the end, I was very happy to have persisted, as there are some real worthy insights here as to how and why, as the subtitle puts it, analogy is the fuel and fire of thinking. It's no small task, thinking about thinking, and a great deal of nonsense has been written about it. Finding out more about how your own brain works, is worth a bit of repetition. We get much more than a bit of repetition here, though.
  
  But, towards the end, after all the threads and examples have been well and truly tied together, every 'i' dotted, every 't' crossed, every point and counterpoint addressed, he does it all again. Through a rather unconvincing dialogue between two characters, just introduced, he goes through nearly the entire book's contents, again, for 20 pages. Agh. Dr. Hofstadter, you are a man who thinks many interesting thoughts. Please allow an editor to guide you in the matter of how many times, and in how many ways, each one needs to be made.
  
  However, in the meantime, I must say that I found the book well worth reading, and I will seek out his next.
• 
  Neven

  I dreaded writing this review.
  
  Douglas Hofstadter is among my favorite writers, and I usually name his 'Gödel, Escher, Bach' as my favorite book of all time. It's a gloriously fun, rambling, clever, surprising, educational, entertaining work that's easy to get lost in and hard to summarize, because it touches on a little bit of everything.
  
  At this point I could make an easy comparison and say that 'Surfaces and Essences' is, in many ways, its polar opposite. You see, this book is far too long, dull, obvious, at times smarmy, and overall not worth your time.
  
  The opening premise is right there in the subtitle: making analogies is the chief mechanism of human thinking and intelligence. Mental acts that don't normally get classified as analogies can, in fact, be understood better in the light of that framework. So far, so good.
  
  What happens then is that the co-authors continue to stretch that premise over what seems like 1,000 pages. (In fact, it is only about 500.) Examples of the same sort of thing are given over and over. (Rather uninteresting examples, too; I don't think any of them would make for good party stories.) They are then dissected through pages and pages of mind-numbingly obvious, often patronizing explanation.
  
  Through it all, chapters and sections are organized almost randomly, with no clear sense of progression or flow. I've always liked Hofstadter's micro-sections (one per page or so) complete with pun-y titles, but here, they're almost parodic. Sometimes, a new section simply continues the previous one. Other times, it changes gears and topics completely.
  
  The writing style is very dull; the jokes are few and corny, the endless clarifications tiring, the tone needlessly argumentative. At one point, the authors complain about "why do so many people refuse to believe that analogy and categorization are the same thing." Really? I've never in my life heard anyone refuse to believe this, let alone "so many people". You probably haven't either. You and I simply haven't given the matter much thought prior to this book. I'm sure the authors have heard this complaint before—they're the ones writing the book. Why drag me, a willful reader, into this fight?
  
  One of the ironies here is that the authors frequently mention just how remarkable it is that we find everyday analogies so unremarkable: we get them instantly, even when the mental work required to unpack them seems significant. But then, why spend so much of this book unpacking them? It simply doesn't seem like a good use of anyone's time to "explain" why we sometimes say things like "turn off the window" instead of "close the window". We all get it; that's kind of the point here. And while being strict and methodical about *one* of these examples is maybe a good starting point for the argument, Hofstadter & Sander just continue to do it over and over and over.
  
  It then comes as a splash of cold water when the final chapters get extremely technical, cavalierly explaining rather advanced concepts from mathematics and physics; not so much to educate the reader about those concepts, but to use them as further examples of thinking as analogy. While I'm not a scientist of any sort, I consider myself fairly well versed in popular science, at least. I can follow most well-written accounts of scientific principles. Since I couldn't really follow these, I'm tempted to diagnose them as not very well written.
  
  The book also overpromises when it says that "analogy is the core of all thinking." That may be true, but no particular part of the book addresses the question of whether other kinds of mental processes are also important. Maybe they're really not, but it would be nice to read that the authors considered it.
  
  Hofstadter has already covered much of this material in 'Fluid Concepts and Creative Analogies', his somewhat technical overview of different AI projects he has worked on. That book was challenging, but I remember emerging from it with a number of fresh insights into computer thinking and human thinking. Similarly, Hofstadter has already talked about the challenges of translation in 'Le Ton Beau de Marot'. Both of these books were, I swear to god, breezier to get through than the few pages devoted to those same subjects here.
  
  This should have been a 20-page paper. I hope the authors—or a more disciplined reviewer than myself—write that paper, so you can get the otherwise solid central idea of this book without all the boring padding around it.
• 
  Martin Cohen

  Here's the short review: Yes, it is rather boring in places.
  
  Here's the long one, because you know, there is a lot in here too.
  
  Behind every word in our language, from nouns such as teapot to connectors such as 'and' or 'but', by way of adjectives and verbs such as 'blue' or 'to paint', "there lurks a blurry richness". Ordinary words don't just have two or three "but an unlimited number of meanings". Why do we use dictionaries then, one might ask? But the fault say Hofstadter and Sander, is with the philosophers, or rather all of them up to one Ludwig Wittgenstein in the 1950s. For them, it was this genius who freed us from the long intellectual legacy of Plato and his notion of heavenly forms for things like, well, chairs and teapots.
  
  "The everyday concepts band, chair, teapot, mess and letter 'A' are very different from specialized notions such as prime number or DNA. The latter also have unimaginably many members, but what is shared by all their members is expressible precisely and unambiguously."
  
  Expressed how? In words? Here's a good word for Scrabble enthusiasts: zeugma. A zeugma is a sentence such as Dickens’ remark about Miss Bolo, that she “went straight home, in a flood of tears and a sedan-chair” where one word (‘in’) is made to do two jobs.
  
  Unfortunately, the first example given, "I'll meet you in five minutes and the garden", is barely comprehensible - at least to me. At least a better one follows a few lines later: 'She restored my painting and my faith in humanity'. But the authors qualify this one as “slightly humorous’.
  
  The book is full of examples illustrating the complexity and fecundity of language - but at 500 or so pages one despairs for a little more selection, a little less repetition. The text is reminiscent not so much of long, learned lectures but of earnest seminars with a white-board on which everyone's suggestions has been carefully written up. In places, the book (metaphorically speaking) lists like the a great ocean liner grounded on the rocks, and for much the same reason - the captain has paid too much attention to entertaining the passengers and too little to navigating the great vessel…
  
  Zeugmas or similar to anaphora, another venerable (medieval) term referring to the way words can 'refer back' to terms used earlier. The fact that the term exists shows how long a history the philosophy of language has too, but little of that of is reflected here, which is a shame.
  
  The book notes briefly that Plato used many analogies, but seem to hold against him that he warned that "likeness is a most slippery tribe". Kant however, is applauded for counting analogy as the wellspring of creativity, and Nietzsche for describing truth as "a mobile army of metaphors". I suspect that Nietzsche was casting his usual aspirations on conventional values tough, and Plato's warning is scarcely incompatible with the authors’ own thesis. They are on stronger ground when they fault the English empiricists, like Locke and Hobbes, the latter of whom, they recall, wrote: "metaphors, and senseless and ambiguous words, are like ignes fatui; and reasoning upon them is wandering amongst innumerable absurdities."
  
  For Hobbes, the mental discourse is true, and the problem comes with 'translation' into words, yet the authors’ own arguments leans this way, except when conceptual imprecision allows for grand creative leaps.
  
  The authors say that they intend to offer "an unconventional viewpoint concerning what thought itself is." The first part of the book aims to show how "concepts designated by a single word are constantly having their boundaries extended by analogies". But whoa! What is the relation between the concept and the ‘thing out there’ - is it one to one? As the word is used more widely, does the concept too cover more ground? If words 'designate' concepts, what use is the (ah) 'concept' of concepts?
  
  The authors look at what they call the development of concepts, for example that of 'Mommy'. This they see being extended over time from being a label attached to a specific human to an allegorical relationship, seen in words like 'motherland', They extend the debate (and this seems to be one of the novel features of the book) from nouns to the many assemblages of language - such as 'after all', 'sour grapes' or 'my Achilles heel'.
  
  As the authors warm to their theme, the look at how "human cognition relies profoundly on our ability to move up or down the ladder of abstraction" (ladder) which they illustrate with a peculiarly parochial example: "For instance, while one is drinking, one will take care to distinguish between one's own glass and that of one's neighbor, but afterwards, when one is placing them in the dishwasher, that distinction will be irrelevant."
  
  Writing about language is all about the quality of examples, Hofstadter and Sander are lucky in that they have enough space in this book to run through many bad examples and still offer (for the diligent reader) enough to count as a profound and thought provoking examination, and their subsequent analysis is usually clear and precise. The dishwasher analogy is an example of what they call caricature analogies "analogies that one dreams up on the spur of the moment in order to convince someone else of an idea in which one believes."
  
  Some of the most interesting analogies though are the scientific ones. The authors argue that "the history of mathematics and physics consists of a series of snowballing analogies". (Snowballing.) For Poincaré, a great thought experimenter as well as a mathematician, analogy is the route towards mathematical discovery.
  
  Naturally, they cite Einstein as a great metaphorical thinker, praising his thought experiments and saying that they helped lead him towards his view of light as particles (rather than waves. But wasn't one of Einstein's key analogies (that he himself credits as leading to his later insights) the analogy of himself as a boy running down a pier with light as a series of waves rolling in from the sea?
  
  Einstein "was driven by an unstoppable desire to seek out profound conceptual similarities, beautiful, hidden analogies." The equation E-mc^2, energy -= mass times the velocity of light squared, is analogous to the rather mundane relationship in mechanics, kinetic energy = mass time velocity squared (albeit divided by two). Equally, as Whorf pointed pout rather earlier than Wittgenstein, words can mislead us. How can light 'weigh' something? Our words (concepts ) confuse us. Light is quintessentially weightless. (Whorf appears here, but only as a brief aside as the 'Sapir-Whorf hypothesis.)
  
  There is a long and substantial discussion of Einstein’s arrival at his theory of the inter-relatedness of mass and energy, but earlier debates in the history of science are sketchily dealt with. The authors explain at length that Galileo’s discovery of moons orbiting Jupiter led to the lowercasing of the Moon and the birth of ‘the concept moon’. Yet, in actual fact the word moon has its roots in the concept of measurement - the moon's monthly cycle being a measure of time. The Moon (capital letter) is derived from this older sense, and so talk of "the concept of moon" being born is misleading, although I think I can guess what they mean - the concept of rocky celestial objects that circle other, larger ones. Yet this is a book about words and analogies - and if at the time Galileo saw discovered Jupiter's moons the Sun was politically obliged to orbit the Earth, the Moon was still supposed to be circling around us, so again what exactly is new in that concept?
  
  The authors condemn attempts to put inverted commas around words, in order to finesse their meanings - but offer themselves a 'typographical convention', whereby when speaking about a word it goes in quote marks but when speaking about a concept it goes in italics. "this is an important distinction, because whereas a word is a sequence of sounds, a set of printed letters, or a chunk of silent inner language, a concept is an abstract pattern in the brain that stands for some regular, recurrent aspect of the world and to which any number of words ... can be attached."
  
  This is mentioned in passing - but if it were really so straightforward, the problems of philosophy would seem to be swept away - and rather in the manner Plato imagines. Concepts seem to have the same role in Hofstadter and Sander's books as 'the Forms' did in the Plato's.
  
  The argument is put that concepts are 'mental categories' made up of many allegorical experiences. Thus a child may call a large, fluffy white dog a sheep, or describe peeling a banana as 'undressing' it instead. These are "semantic approximations', the "concepts silently hidden behind these words will continue to develop in the minds of all these children".
  
  "Could it in fact be the case that the tiny child's act of calling an everyday object by its standard name is a close cousin of the genius's act of creating a new concept that revolutionize human life?"
  
  However, later we are told that the "classical view of categories" is now perceived as a dead end (a dead end) and so contemporary psychologists "have tackled the challenge of making the very blurriness and vagueness of categories into a precise science". Out go precise membership criteria and in come instead the notion of prototypes. What are these? They are generic mental entities "found in long term memory" which summarises all someone's experiences, which consists of a full set of exemplars encountered over a lifetime, and finally, we are told that "certain regions of the brain that were once stimulated by the closest experiences to the current stimulus" are re-stimulated.
• 
  Bart Jr.

  Surfaces and Essences by Douglas Hofstadter & Emmanuel Sander 12/08/2014
  
  
  
  Surfaces and Essences by Douglas Hofstadter and Emmanuel Sander
  
  Fantastic book, full of interesting insights about analogies. And analogies about insights. I always try to distill a book like this into what I consider the key points.
  
  The first of those, in my mind, was the definition of intelligence offered. Intelligence is the ability to rapidly and reliably see the crux, the gist of a situation. I thought about this. Was this the single most crucial element of intelligence? Just asking the question is a clue to the answer.
  
  Naturally enough, I went to the bookshelf and pulled out Jeff Hawkins' book On Intelligence, to see if I could gain some insight. His take on the mind is that it is a memory-prediction device. It was almost as if these guys were climbing up different sides of the same mountain. Because if you really have the ability to see the crux of a situation, wouldn't you be in the best shape to make accurate predictions? And he also makes the point that the mind uses memories as analogs to predict, while Hofstadter and Sander think that the mind uses the analogies offered up by memory to categorize. In fact they think that categorization and analogy are really two sides of the same coin.
  
  So while the focus of the two books is different, Hawkins' being on the physical structure of the brain and how it works to create intelligence, and the focus in Surfaces and Essences being analogies and how we use them to categorize, they both seem to agree also on the fundamental role that memory, or analogy, plays in thought and categorization. These two approaches actually complement each other.
  
  So, I came away with a definition of intelligence that I thought did find the crux of what intelligence is. And also the fundamental importance of analogies in thought and categorization. The book was also full of interesting analogies of all description and variety.
  
  I also love the use of the Copycat domain, which gives an amazing variety of peeks into the creative slippage of concepts and how complicated even the simplest can be. Every page of this book contains insights into thought and language. I would like to see Mr. Hofstadter examine symbolization itself more deeply, in speech, writing, computer use, etc. but I suppose that would be another book.
  
  
• 
  Dave Peticolas

  Fantastic. This is Hofstadter's latest thinking about thinking and, as usual, he has some enormously interesting things to say and a delightful way of saying them. Highly recommended.
• 
  Aloha

  Fabulous as only Hofstadter can make it.
• 
  Sylvia

  This book has a nice thesis (essentially all thought is merely the construction of analogies), and I enjoyed reading it. I felt that much of it was indulgent--full of dozens and dozens of examples where a few would have sufficed, and paced rather glacially, but it was quite readable and often pleasant.
  
  The last few chapters, though, which highlight analogies in the fields of physics and mathematics were AMAZING. The authors step you through a series of analogies that lived in the minds of the physicists who first investigated the atom, and it honestly took my breath away because I felt like I could finally appreciate the conceptual leaps which Einstein made. Suddenly, hundreds of pages into this book, I was compelled to talk about it obsessively with other people and reread pages over and over again.
  
  Next, there was a segment on the hidden mental analogies which guide our approach to math problems (e.g. multiplication as repeated addition, division as portioning out). As a math teacher this was really interesting and informative, and I will return to look at it again.
  
  Overall, a good book--but DEFINITELY read the last few chapters if you are math/science-minded, they are the best part.
• 
  Elizabeth

  When I started listening to the audiobook version I thought this was going to be a solid 4 stars. About half way through I had knocked it down to a three for the same reason many people have cited - the repetitive nature of the examples was becoming overwhelming. By the time I reached the last chapter, my final assessment hovers somewhere between a two and three. I felt that the last chapter - a manufactured debate about the thesis that was more than thoroughly covered up to that point was complete overkill.
  
  I think it is an interesting premise and there are a lot of thoughtful passages and well articulated points, but brief it is not. You'll find six examples given for every one required to make a compelling argument. If this book was edited down to the most salient points and examples, I would have given it a four or five star rating.
• 
  Gary Beauregard Bottomley

  I didn't listen to the whole book. I listened to about 9 hours of it. The book is clearly for lovers of words of which I'm not. I do like the authors overriding theme that we think by categorization through analogy. I just didn't want to sit through a countless stream of analogies and word play examples.
  
  Some people (especially lovers of words) will love the book. I just prefer less examples and more facts.
  
  (I bought this book because I absolutely loved the senior author's book, "Godel, Escher and Bach", and suspected that this word book would not be for me, but was willing to give it a try. I'm only writing this review to warn people who prefer science and mathematics type books that this book might not be suitable for you).
• 
  Jafar

  I can't judge whether the authors manage to make a convincing case for analogy-making being the basis of human intelligence. That's beyond my qualifications, and moreover, I skimmed through the middle parts of the book. But there's such an amazing trove of interesting facts in this books, especially on the linguistic front, that the book is worth reading regardless of the its grand claim. There's a long and great chapter at the end about how Einstein used analogies to come up with his special and general relativity theories. That chapter will give you a completely different but better understanding of relativity than the ones that you've seen in other books and are used to.
• 
  Baal Of

  squirrely briggates
  
  unseedy
  
  square braces
  
• 
  Steve Abreu

  A long, but very enjoyable book about the role of analogy in thinking - from mundane everyday situations all the way to creative strikes of genius by Einstein and the likes. Many gems along the way, along with big take-home messages like analogy being one and the same as categorization. Very convincing, it will take some months to fully appreciate (or reject?) their ideas. Not to mention the question, what does this mean for building AI? And, if analogy-making really is the core of cognition, what are the neural mechanisms for it? Is analogy-making directly connected to the predictive memory and reference frames that Hawkins advocates as the core of intelligence? Too many questions..
• 
  Dan Falk

  My first grade-school reader was titled A Duck is a Duck. (Stay with me on this one, folks.) Now, by first grade I had learned a few things about ducks: They have feathers and a beak; they can swim and fly. No doubt I had observed that pigeons are sort of like ducks (they have everything except the swimming part); ostriches, meanwhile, don’t look much like ducks or pigeons – they can’t even fly – but they’re close enough to ducks and pigeons to deserve the label “birds.” Of course, airplanes have wings and can fly – they’re not birds, but perhaps we can place them in a larger category of “things that fly”…
  
  Knowing what makes a duck a bird and what makes a plane not-a-bird may not seem like very profound mental feats – but Douglas Hoftsadter and Emmanuel Sander see such cognitive connections as part of an extraordinarily profound process. In fact, they see the task of putting things into categories, and the related task of drawing analogies, as the very essence of thought. (It is not giving too much away to reveal that, toward the end of the book, they argue that these two tasks are intimately related.)
  
  The jacket copy for their hefty new book proclaims it as “an ambitious new theory of how the mind works” – a bold claim which, if it came from lesser minds, would call for extreme skepticism. But Douglas Hofstadter, a cognitive scientist and computer scientist at the University of Indiana, is widely regarded as an intellectual heavyweight, and when he says that “analogy-making defines each instant of thought, and is in fact the driving force behind all thought,” one feels compelled to at least hear him out. After all, his first book, Gödel, Escher, Bach (1979), won the Pulitzer Prize, and a string of follow-ups have established him as one of the most creative scientific thinkers of the last few decades, and a highly skilled conveyor of complex ideas. His younger co-author, Emmanuel Sander, is a psychologist at the University of Paris.
  
  They begin inside the mind of an infant: A child’s first thought, the authors suggest, is “mommy”; the child identifies this concept with a specific physical entity, namely the caregiver who provides food and comfort. Later she learns that other children also have mommies, and, eventually, that animals have mothers too; and also that “father” is similar to “mother” but not quite the same thing. This impulse to categorize and compare stays with us, for “analogy is our perennial dancing partner.”
  For scientists, categorization and analogy-making are even more crucial. Consider Galileo’s first observations with the newly-invented telescope. Before Galileo, there was only one “moon,” namely the earth’s moon; to speak of “moons” was simply gibberish. But when the Italian scientist observed four small, star-like bodies seemingly revolving around the planet Jupiter, he was forced to do a re-think: Any celestial body that revolved around a larger one might very well deserve the label “moon.” Another example is the broadening of the category “number” to include quantities less than zero. Today any fifth-grader can add and subtract negative numbers, but in 16th century Italy they were simply unimagined. When negative numbers were finally granted the same status as positive numbers, a whole slew of vexing problems suddenly yielded simple (or at least manageable) answers. The book’s most persuasive section involves the physics of the early 20th century, as the authors focus on the thought processes of Albert Einstein, who they label an “analogizer extraordinaire.”
  
  Be prepared to become hyper-conscious of the myriad of analogies one makes every moment of every day. (“He’s using this Starbucks as his office”; “City Hall is a circus”; “This book is a bit like a doorstop.”) And be prepared for overkill: The question of what is and what is not a “sandwich” goes on for several pages. And a list of what the authors call “caricature analogies” runs through 24 examples, at which point we are told: “Such a list could be extended forever.” No kidding.
  
  The dangers of putting things into categories gets only a brief mention. Recognizing that kangaroos and koalas are related means you know your marsupials; but linking “young black man” and “criminal” simply makes you a racist. And think how much is at stake when a politician or a news outlet classifies a violent attack as “gang-related” or an “isolated incident” or an “act of terrorism.” Hofstadter and Sander do not address these questions. They only briefly mention the problem of stereotypes, which “have a bad reputation” but are also “crucial to our survival.” They later note that stereotypes “are a frequent source of deeply erroneous categorizations” – a warning that comes three pages from the end of the book.
  
  Some of the arguments in Surfaces and Essences feel a little dated – after all, the power of metaphor is hardly news, and Einstein’s thought processes have been probed up, down, and sideways. Meanwhile, some current topics, such as neuroscience, are simply bypassed. The end result is a book that is ambitious and provocative, though unfortunately lacking the originality and spark of Gödel, Escher, Bach. (Adapted from a review I wrote for The Globe and Mail.)
• 
  Chris

  What does it mean to understand a concept? What does it mean to communicate or teach someone a concept? What does it mean to know a concept or to not know a concept or to conceptualise an idea? What role does language play in that process?
  
  Two masters of their fields, Hofstadter being of particular note to me, due to his notable advances in artificial intelligence, argue how analogy is the basis of all thought.
  
  The book is a slow, highly-detailed and careful analysis of language, abstraction, categorisation, and how analogy relates to it all, aiming to help us understand the broader system of how thoughts and ideas are learned, associated, communicated and conceptualised.
  
  The thesis is not difficult to understand: We think by conceptualising our experiences within socio-cultural-linguistic boundaries. This process of concept development, encoding and retrieval is done through continuous analogy-making (categorisation). The analogy, is therefore, the medium through which concepts are stored and interact with one another through reference and comparison. Analogy is both a component of, and following process of, the extrapolation of the "essence" of the concept, which determines its relevance to other concepts. Analogy also involves comparison of categories through abstraction. Thus, it is through this process of concept categorisation and abstraction we are able to build mental models (simplifications) of the infinitely complex universe around us.
  
  Despite, its simplicity, however, this is one of the most important and life-changing books I've ever read. It is about as meta as you can possibly get. It's the essence of meta, pushing the boundaries of metacognitive awareness into a hall-of-mirrors feedback loop that breaks your mind. When you emerge from the other side, you have undergone a metamorphisis, and you will quickly see that you have an understanding of the universe you did not previously have.
  
  For example, by understanding the role analogy formation (conceptual categorisation) takes place, you will be able to better understand, with certain methodological precision and confidence, why someone just doesn't get you, get what you're talking about, get your idea, or misunderstands the underlying premise of what you're sure they got. And why they did. You'll be able to look at otherwise estoric concepts that are themselves based on conceptualisations of concepts, such as academic research, to better infer what it actually is. Although this is not discussed in the book. It doesn't need to be, as you'll be able slide through the right level of abstraction to draw commonalities between core research ideas such as theory, level and unit of analysis, and ontological orientation, and the notion of analogy.
  
  To address the other reviews, many of which demonstrate the inability to distinguish between the frustration the reviewer feels for the...slow...writing style and the quality of the underlying thesis, they are correct that this could have been condensed considerably. Did it need long lists and highly detailed analyses of common words, idioms, and proverbs? No. They would have been better as an appendix. Would it have increased the quality of the book and refined the thesis if the number of examples were cut by a factor of five or ten? Yes, most certainly. However, the number of examples fuel thought and slow down the reading, which can, if you allow it, give time for increased reflection. I'm also not convinced it really needed the foray into Einsteinian physics in the last chapter, which I struggled to follow, nor the epidialogue, which I must admit, I found too tedious to care much about. But, some people might find those chapters very insightful.
  
  This is a book that requires patience, but will provide, in exchange for that patience, a superpower of seeing the world in an entirely new light.
  
  
  Here are just a few key points:
  
  Language
  • Semantic space, such as nouns, phrases and proverbs are analogies formed through inference of the universe. The word "bird," for example, is analagous of a cluster of multidimensional inferences of what a bird is, which represent the schemas that are established through experience, eg. a bird has feathers and flies. But is loose in essence (a plucked bird that cannot fly is still a bird).
  • Similarly, the semantic space behind even mundane phrases extends far beyond it's immediate meaning, it is a representation of a number of future events or concepts that are multi-layered and highly contextualised. Eg. 'What's new?' Is not just asking about recent news, but also denotes the desire to express freindship and care.
  • Word and phrasal analogies are not aligned uniformly across cultures and languages, and the associated analogies embedded within the local semantic space can differ hugely. Although, there are a core of the conceptual space shared within cultural and national linguistic boundaries. Hence, there is both distinct and common "lexical coverage" unique to each language.
  • As the partitioning of a category is an infinitely fluid process of finiteness and abstraction, it is impossible for a word to exist to represent every common category. One way of overcoming this is to reuse words in multiple categories and contexts, yet the boundaries of application must be finite to maintain distinctiveness.
  • Another solution is to combine words, and by doing so create analogies that are representative of new categories that are distinct from their constituent parts. This may be seen as indivisible compound nouns, eg. Nightcap, or a phrasal verb, eg. Look down on, or contextual sentences, eg. 'Are you out of your mind?' or an idiom, acronym, proverb, or common fables.
  • The inferences behind analogies, even ones representing a complex context, such as reducing congitive dissonance when facing defeat through distortion of original intention, is analogised into a transferable essence: "sour-grapes-ness," which fits a large number of other contexts, when expressed as a common category through semantic space, such as via proverbs.
  • The essence of any context is automatically related to previous experience through a complex but spontaneous encoding of any of each of the salient details of an event, which are distilled to fit a conceptual skeleton that can be recognised, also spontaneously, in entirely different experiences.
  • Words are transferable to different levels of abstraction eg. Trucks are a form of cars, cars a form of motorised vehicle...etc... This process of "marking," where the membership category of a word changes across levels of abstraction is a fundamental component of communication and conceptualisation. The word's embedded ambiguities allow for adaptation of context, which permits commonalities to be drawn between radically different contexts.
  • Metaphors often exploit the abstract category of a concept. Eg. "He's a pig" is not an ad hoc association but utilises the broader abstract category of pig, rather than the concrete category of a pig being a farm animal, where pig-ness implies messiness, sloppiness or dirtiness, which can then be associated to behaviour or attitude.
  • Metaphors can also be extended so that abstract categories become distinct from the underlying concrete category. Eg. "She's a bitch" has no association with the essence of being a female dog, but has formed it's own distinct associations that have no connection within the mind to "female-dog-ness."
  • Metaphors that use new ad hoc concrete categories require a shared understanding of the inherent essence so that it can be extended into abstract form.
  
  Learning
  • Expertise involves expanding both horizontally, by forming a larger number of concrete categories (eg. Dog breeds), and vertically via hierarchical abstraction, shifting between concrete, lesser salient features and more abstract categories (eg. Golden Retriever-dog-mammal-animal). It also involves analogies from points of commonality and differences across categories at mutliple levels of abstraction (category interconnections).
  • This unconscious structuring of mental categories are our mental models: our way of simplifying the infinitely complex universe comprised of infinitely comparable and divisible categorisations of possible concepts at infinite levels of abstraction.
  • The level of abstraction of the finite mesh of categories define the form of expertise: a higher number of concrete categories allows for higher precision of expertise, while a higher abstraction of categories allow for depth of knowledge and more cross-category inferences.
  • This stage of domain knowledge development is one of evolution from categorisation of surface-level salient features to deeper, more abstract categorisations. When a sufficiently abstract understanding of a domain has been achieved, the most salient domain features are inherently more abstract.
  • As expertise grows through abstraction, new contexts (those not encountered during learning) can be compared to the salient features of higher abstract categories via analogy to extend domain knowledge further.
  • Innovation often requires adopting a more abstract viewpoint to repurpose an item that had been previously categorised in a more finite context to recognise its potentially more general purpose. Eg. Shoe laces as chord for securing items.
  • Analogy may lead to inappropriate or incorrect connections in multiple ways. We can misgroup words, concepts and experiences through the salient properties of the concepts or experiences. This leads to speech errors and incorrect judgements due to the inability to correctly distinguish between categories, and illogical induction during decision-making due to misrepresentation of the perceived likelihood of reoccurrence.
  • Invisible interfaces: by designing objects (or software) in a way that analogises everyday familar objects in a goal-oriented manner, you can create interfaces in a way that the user doesn't realise they're interacting with an analogy or even an interface. Eg. Files/folders on a computer are so intuitive that we don’t recognise that they're analogies for interacting with locations in memory.
  • Naive analogies: we learn the "nature" of concepts by analogising them to simple, naive analogies. Eg. Division as a form of "sharing." This can remain permanently and cause conflicts when requiring a different analogy (such as trying to divide by a decimal of between 0-1) as one cannot "share" something and get a larger number as a result. Here "measurement" would be a better analogy ("how many times does .5 go into 4?"). Hence, naive analogies can be completely wrong, preventing understanding and learning.
  
  
• 
  Samoht Neerg

  Have only read first few chapters, great ideas but feels too indulgently long and repetitive. I’m sure there’s nuance to each example but not enough here to keep me interested. Maybe worth reading in full if you’re seriously studying language, cog-sci, NLP or linguistics but I’m not so got bored as I couldn’t relate more than high-level concepts to everyday life. Might return to this one day. There’s a video online where Hof talks about “analogy as the core of cognition”.. Interesting ideas and some affinity with George Lakoff’s Metaphors we live by.
• 
  Sameer Vasta

  If the goal of reading a book is to learn something about yourself or about life itself, then the most important thing I learned from reading
  Surfaces and Essences: Analogy as the Fuel and Fire of Thinking is to never ignore the sticky notes another reader has left in the front of the book.
  
  I picked up
  Surfaces and Essences: Analogy as the Fuel and Fire of Thinking from the library, as I do most books, and upon opening it, I discovered someone named Libby had left a note inside to be read before beginning the book. Being slightly obstinate, I ignored her pleas to stop and read her notes before rifling through the book.
  
  The full text of her note is below:
  
  

  Stop! Before you go any farther — before you dedicate a minute, an hour, a day (or days) on this book, there is something you need to know: it’s not worth your time. There. I said it. It took a lot of guts, trust me. I can express to you in about 2 sticky notes the entire book:
  
  An analogy is drawing an inference or comparison between two objects or concepts. it’s the rhetorical equivalent of a metaphor. Or: A is to B, as C is to D.
  
  That is all you need to know. Sure, there are great examples throughout S&E (mostly about two men going out for “coffee” but not ordering coffee) but Hofstadter always goes one step too far. Please read these section on the sandwich for a perfect example of what I’m talking about (pg 214). If you can stomach that, then you might make it through the book.
  
  Many of the ideas are hijacked from others & the organization is pretty weak.
  
  That being said, you will read what you want to read, my opinion be damned. But don’t say I didn’t warn you!
  
  Libby

  
  
  She wasn’t wrong.
  Surfaces and Essences: Analogy as the Fuel and Fire of Thinking took almost 700 pages to explain what a basic analogy and a basic equivalency was — all things many of learn in our freshman year of college. There are probably more concise and entertaining articles online that explain it further; investing several hours into this tome is unnecessary.
  
  There are some humorous and some interesting passages, and some of the examples help make comprehension easier, but overall, the key learning from
  Surfaces and Essences: Analogy as the Fuel and Fire of Thinking is something that can be quickly expressed in a much shorter, and more captivating, piece of writing about rhetoric.
  
  I did learn one key thing: if I ever come across another book with a note from Libby, I’ll be sure to read it first and heed her advice.
  
  (
  Full review on I Tell Stories.)
• 
  Richard Thompson

  This book was a huge disappointment. It has a promising premise — that categorization and analogies are basically the same thing and that they are the key building blocks in human thinking. And I found it about 75% convincing, but the presentation is awful. The book is about five times longer than it needs to be and makes the same points over and over again, using a seemingly never ending stream of examples far more extensive than is needed to make the point. And then the authors try to categorize and define different types of analogies, which by their own logic can’t really be done with any degree of accuracy.
  
  I was also concerned with a couple of key parts of the argument on a substantive level. First, at several points in the book, the authors argue that the process of analogy making is shaped by the forms of language, in effect that language shapes thought. But this just isn’t so, as discussed in great detail in McWhorter’s book, “The Language Hoax”. They do acknowledge that the more extreme aspects of Whorf’s theory that language influences thought may not be correct, but they still go way too far in embracing this discredited theory in support of their thesis. Then they go too far again in asserting that analogy making is the basis for scientific advancement, including a detailed discussion of how Einstein used analogies in developing the theory of relativity. I am prepared to accept the idea that analogy may have been an important tool in Einstein’s toolbox, but I think only one of many. These guys should read Paul Feyerabend’s wonderful book “Against Method”, in which he makes a powerful case for the idea that there is no single method that is or can be the basis for scientific advancement. Any attempt at boiling down the ways of science to a single tool or method is mistake, regardless of the chosen candidate method.
  
  Seeing these mistakes in key parts of the reasoning cast doubts for me on the validity of the overall thesis. If I can see two errors in the reasoning, how many others are there that I did not see? In the end I still found the book to be persuasive, but deeply flawed.
• 
  Nick Klagge

  Unfortunately, although I am a huge fan of Hofstadter's "Godel, Escher, Bach," I found this book to be much less impressive. S&E lacks the engaging format of GEB, with only one short dialog at the end (rather than interspersed throughout, as in the earlier book). In addition, I would say that S&E could have been significantly shorter, and probably would have been edited more aggressively if Hofstadter weren't such a big name. The book often goes on for multiple pages illustrating a topic where a few short examples would suffice. A final thing that frustrated me was that the authors generally did not clearly address where they stood in relation to the existing literature on cognition. As a non-expert in the field, it was difficult for me to tell how new or radical their argument was.
  
  All that said, I found the book quite convincing, and Hofstadter is generally an enjoyable author to read (if you like his sort of humor, which I do). My main takeaway from the book, however, was a curiosity about an application that the authors did not address directly, or really at all: the role of analogy in the formation of utility functions (or preferences, choices, however you want to characterize it). So many of our choices involve options that we have never experienced directly, meaning that we have to evaluate some or all options based on some type of analogy. Even for things we have experienced in the past, new experiences of them may not really be "the same" as our previous experience, which means that the obvious analogy may be misleading. Finally, the distinction drawn by Kahneman between the "remembering self" and the "experiencing self" may muddy these choice-analogies even further.
• 
  Alex Weird

  As a linguistics student fascinated by cognition, I love the book for all the reasons others hate it.
  
  The long lists of examples are joyous to pore through, providing a dense and diverse corpus of illustrative ammo to draw from in your own conversations and projects.
  
  This is true in general of the book's overall length and repetition: it's just richness and comprehensiveness. Instead of a quick chug of meal replacement, the 500-so pages are a thicc, delicious stew for you to savour and truly digest.
  
  Had the book only been 200-300 pages, there are some powerful ideas and observations of my own that never would have occurred to me. The length and reiteration gives you the time and reinforcement to deeply cognize the material and cultivate your own thoughts on the subject.
  
  A shorter book would have given you the *feeling* of understanding, but true knowledge demands the thorough, balls-deep approach Sanders and Hofstadter take.
  
  This was one of the top 3 reading experiences of my life, next to Gravity's Rainbow and of course, GEB.
• 
  Rachel Moyes

  It started out so good. Then it got so long and boring. Then there was a whole chapter about scientific discoveries that felt very self-indulgent. Like, you know Hofstadter just wanted to include that because he thought it was cool. It was overly technical and could have been its own book.
  
  Then the ending was bizarre and completely unnecessary and very self-indulgent again. It's a dialogue between two people used to sum up the book. One, Katie, is arguing that categorization is the essence of all thinking, while Anna argues (correctly) that analogy is. Katie is just there to raise straw man arguments that Anna easily shoots down. The worst part was when Anna REFERENCED THE BOOK in her arguments, like, "Well, if you'll remember in chapter 4, Katie..." Breaking the fourth wall in the worst way. And we didn't need a summary. Everything had been so summarized in so much detail during the chapters. It felt patronizing--of course we get the categorization and analogy making are the same thing! We've already read 500 pages of this book!
• 
  Michael Dubakov

  This book is quite cool. It is about thinking and the authors main idea is that all our thinking processes based on analogies. They provides 530 pages of arguments in defence of this idea. While it looks convincing, the book is quite verbose. I believe it can be easily cut to 300 pages without any problems. Information-to-inc ratio could be easily improved, that is why I gave it 4 stars.
  
  Last third of the book is brilliant though. There are very cool examples from math and physics I really enjoyed. Most people have bad intuition about multiplication and division, and modern education is a root of these problems.
• 
  Shane Mcloughlin

  This book was a great, easy to read, book about analogical cognition written by two cognitive scientists. While cognitive science doesn't provide an adequate model of what analogical cognition is and how it might be trained / manipulated, it does a great job of stressing its importance in the first place. The narrative that makes this book easy to understand would never be found in one of the articles that actually provides an empirical model of analogical responding. In this case, that's a great thing. A very well written book.
• 
  Mickey Kawick

  It is a terrible book. The author uses every comparison in the world as an analogy... this sounds like that... must be an analogy... this word is combined with another word... must be an analogy.
  
  Boring, repetitive, and I find that a man whose only tool is analogy sees the entire world as an analogy.
  
  The world is filled with more subtlety than that. Many other people here have read the same book and found it similarly dull, uninsightful, and childish. Save your reading time for a good book.
• 
  Mani

  This is now officially one of my favorite books. Towards the end it got tedious and I suspect that it's one of those books that just causes one to just reach in there and jostle one's own thoughts. Also I read it right after Dennett's Intuition Pumps- if that affected my state of mind. I would recommend it if you're into thinking about words, thoughts, and psychophysics.
• 
  Rogers George

  As I read, I felt that I already knew everything in the book--until I got to the last chapter, which contains the clearest explanation of why E equals mc squared that I have ever read. I felt that I knew the material intuitively, but the book organizes it, discussing and codifying how we think in analogies--in great detail, and with Hofstadter's trademark subtle humor.