Analogy as the Core of Cognition

Joe Becker's 1975 article 'The Phrasal Lexicon' argued that the mental lexicon is far richer than a set of words; it contains phrases, idioms, and other linguistic chunks that are stored and retrieved as holistic units, not constructed compositionally.

“I got some of these ideas. In fact, I was deeply influenced by an article that I read in around 1975 by Joe Becker, who's here tonight, called the phrasal lexicon. And Joe pointed out that what we have in our head is far, far richer than a set of words when we talk about our mental lexicon...I can't possibly do justice to his delightful and stimulating article. I mean, very very inspiring article.”

Human minds have an unlimited capacity for chunking: lower-level concepts combine into higher-level conceptual units that become semi-invisible once chunked, and multiple layers of chunking create hierarchically organized concept networks where internal components fade from awareness.

“Our minds are constructed Oh, this is no not new to anybody. That that with a an unlimited quality for chunking. By that, what I mean is that primordial concepts in some interrelationship become uh a larger unit, a larger conceptual unit like, you know, a family or something like that...we build our concepts by taking several concepts, putting them together, putting a membrane around them, and then, sort of miraculously, the internal components sort of disappear. And we're left with just this new concept, which is kind of like a black box...these things become semi-visible. They're sort of lurking in the background. But, if it's structured at many levels of a hierarchy, the concepts inside the concepts inside the concepts are certainly just about invisible.”

The concept of 'mess' (as in a visually chaotic or disorganized state) cannot be explained by feature detectors; there are no specific neurons or feature detectors that fire when seeing a mess.

“If we're talking about visual categories, what are the features of a mess? I would like to know which neurons, which feature detectors in my brain fire when I see a mess.”

If strokes of genius, personal insights, political decisions, dinner table conversations, me-too comments, random remindings, instantaneous categorization, and blends of all sorts are all made of analogies, then might not all of cognition also be made of analogies?—the cognition core hypothesis.

“if strokes of genius are made of analogies and personal insights and political decisions and dinner table conversations and me too comments, which I didn't cover unfortunately, and random remindings and instantaneous categorization and blends of all sorts are all made of analogies, might not all of cognition also be made of analogies? That's the cognition core hypothesis. Remember, you heard it first here.”

Many analogies are reminding events that happen continuously and involuntarily; analogies are not there to serve purposes but simply happen, with evolution being the only purposeful driver behind them.

“a lot of analogies are reminding events...a lot of people think analogies are there to serve purposes. They're not there to serve purposes at all. They just happen. That's all. There's no purpose. There's no purpose at all. They just happen. Uh sure, they they're in a certain sense, they're serving evolution, but that's about the only purpose that there is”

The Gulf Stream is a warm ocean current that flows north toward the British Isles and Norway; England blocks the Gulf Stream's flow, casting a 'shadow' of the Gulf Stream such that southern Norway (which is blocked by England) is colder than northern Norway (which receives the Gulf Stream's warmth).

“the Gulf Stream the Gulf Stream uh is warm...it goes north...where it's cold is southern Norway...Because the Gulf Stream is blocked by England...There's the Gulf Stream which is supposed to be red here and uh and it gets blocked by England. And so the shadow of England is on the coast of southern Norway and and the Gulf Stream hits the up upper part of Norway. So we have a shadow thrown by England uh of the of the of the Gulf Stream.”

Analogical reasoning is conventionally treated as a minor component of cognition relegated to one small zone, but analogy should instead be understood as the central mechanism that connects everything—analogous to an interstate freeway system that links all of cognition together, not a peripheral Delaware.

“in cognitive science, the cognitive science conferences and journals and so forth, they always have they divide cognition up into all sorts of different pieces, and one of the pieces is always called analogical reasoning...they sort of is is the Delaware of cognition...Analogy is the interstate freeway system of cognition. It is not one little tiny zone somewhere off in the side.”

Two native English speakers conversing in a restaurant neither noticed when one said 'easy-go-lucky' (blending 'easy-going' and 'happy-go-lucky'), demonstrating that most blending events pass completely undetected by both speaker and listener.

“native speakers talking two of them talking in a restaurant. Neither of them noticed this comment. I mean, it just went by. Nobody heard anything. Uh he was a real easy-go-lucky guy. Easy going and happy-go-lucky.”

Categorization is the fundamental game of cognition, and analogy is the mechanism that creates and enables categorization by allowing the mind to perceive what something is and grasp its essence.

“Categorization is the name of the cognition game, but analogy is the mechanism that creates or that allows categorization to happen. By categorization, I mean, uh deciding what something is, what the essence of something is.”

Repeated analogies expand and generalize concepts over time; by seeing new instances that share essence with existing concepts, we broaden the meaning and applicability of those concepts, both for individuals and for culture as a whole.

“here is uh just you know, this simple idea that uh repeated analogies expand concepts...Concepts expand over our lifetime and they expand for each one of us and they expand for our culture.”

Categories are not restricted to nouns and visual objects; words of all parts of speech—adverbs, prepositions, pronouns, verbs—are categories evoked by appropriate circumstances; most categories are not visual or sensory at all, and theories of categorization that focus on visual objects miss where the real action is.

“I am not favoring nouns. I want you to understand that when I talk about categories, I do not think that categories are always visual things like you know, overhead projector or shoe or you know, even audience. Categories can be things like as I said, please...There are certain circumstances where that word is evoked...It's a please circumstance...words of all sorts, whether they're adverbs or prepositions or pronouns or verbs or nouns, they are all categories and they're all equally good and most of them are not visual. Most of them a lot of them are not sensory whatsoever. And theories of of categorization mostly are involved with visual categories, visual objects. It's so mistaken. That is not where the action is.”

Pluralization is the process by which a proper name or unique entity becomes a common noun and gets applied to multiple instances with shared essence; examples include calling multiple intellectuals 'young Einsteins', referring to cities as 'the Paris of' a region, or using 'Mecca' as a generic term for any destination.

“thing I call pluralization...There may be two or three young Einsteins among us tonight...Milton Babbitt is no Mozart...the next Catcher in the Rye or the Soviet Union's Vietnam, the Paris of the Middle East...Jerusalem being described as a Mecca for tourists...Quisling was a Norwegian uh foreign minister...it became a lower case word for basically a traitor...Galileo...called those moons. That was an a magnificent leap that pluralized something that nobody had ever ever considered to be other than one unique thing in the world.”

Hofstadter's assistant Helga said 'Every morning I took a capsy ride,' blending 'cab' and 'taxi' into 'capsy,' which reveals subterranean competition between two synonymous words for the same vehicle.

“my administrative assistant, Helga, said, 'Every morning I took a capsy ride.' Beautiful. Beautiful.”

When Hofstadter said 'finding finding' and prolonged the F sound, he was experiencing subterranean competition between 'find' and 'figure out,' which became audible only because he hesitated to articulate a sound that he consciously registered as slightly too lengthy.

“I was having trouble finding finding. The F was a little too lengthy. And I said did anybody hear that? Nobody even heard it... I know there was competition in my head and I happen to know what it was. At least part of it. I don't know the whole story, but at least I know one of the words was find and the other was figure out. And I couldn't quite decide. Now how much else was going on inside inside my head I don't know.”

The concept of 'airline hub' is built from many nested concepts including: ball→roll→wheel→spokes→hub→centrality; vehicle concepts; company concepts; trip→legs→routes; time→appointment→timetable→network; network→node; get; buy→money→economizing→downsizing; city; airport.

“I've put here things like ball goes to roll, roll to wheel, wheel gives you spokes and hub, and gives you the notion of centrality...bike is a vehicle, a bike has wheels, car is a vehicle, a bus is a vehicle...Then we have the idea of a trip, and then a trip can be broken up into legs, and legs can be sort of formalized into routes...notion is abstract notion of network goes to node...when we say, you know, Denver is United hub, do you think about all this stuff? It's you don't think at all. I mean it's it's you you you know this as a chunked concept.”

There is no fundamental difference between a single memory trace and a category or concept; a single instance can function as a concept, and conversely, a concept with many blurred instances is structurally equivalent to what we'd call a category.

“The the basic idea here is there is no fundamental difference between a single memory trace and a category by which I mean a concept. No difference category concept, same thing. Okay, that's a very important idea of this talk...There's a no fundamental difference between seeing something as a category category member and seeing something as analogous to something else. In one case, something evokes a concept which has one instance, and another case, something that evokes a concept which has many instances and which has blurred together.”

Phrases like 'There but for the grace of God go I', 'Pretty please with sugar on top', 'Been there, done that', and 'Put your money where your mouth is' are mental categories evoked by situational contexts; they function as situation labels and category names.

“phrases...There but for the grace of God go I. Pretty please with sugar on top. I think that's a nice one. Been there, done that...Put your money where your mouth is...these are evoked by situations. These are categories. These are categories. And I want you to understand that all these things are mental categories.”

Analogy is the perception of common essence between two things, where the essence is subjectively perceived based on the particular frame of mind one is in at the time, rather than an abstract philosophical property; analogies occur inside the head as connections between mental representations, which are then projected onto the external world.

“Analogy making is the perception of common essence between two things...things don't have essences, but what I mean I I'm not a, you know, I'm not talking about some kind of abstract glowing philosophical essence. I'm talking about the essence that you perceive at the particular time in the frame of mind that you happen to be in...analogies happen inside your head, so that they're they're they're connections between two mental representations. They're connections between things inside your head. Which we project to the outside world.”

Most analogies happen so quickly and transiently that they would pass completely unnoticed and leave no effect on the world at all unless someone is paying deliberate attention to the mechanisms of their own thinking.

“this event inside my head happens so quickly and so fleetingly that I would not have even noticed it if I weren't kind of an inveterate observer of my own thinking...it would have disappeared in a flash and it to have no to leave no effect on the world at all.”

Hofstadter's colleague claimed that categorization is 'a matter of feature detectors', but Hofstadter disagrees, saying they are 'not on the same planet', because categorization at the level of abstract, non-sensory concepts (like 'probably', 'mess', 'well') cannot possibly rely on feature detectors.

“one of my colleagues who I respect very greatly said that he thought that categorization at this kind of level was very very straightforward. It's just a matter of feature detectors. And I think, 'Oh my goodness, what we are not on from the same planet.' I don't think categorization at that level is anything to do with feature detectors.”

Roger Schank pointed out that proverbs are situation labels; sometimes proverbs are further compressed into single-word labels, as in 'sour grapes' (lower case) which evokes the situation of someone rationalizing away a desire for something unattainable.

“this is something that Roger Schank pointed out. Proverbs are situation labels...sometimes proverbs are boiled down into just a single word like sour grapes. I know that sounds like two words, but I see it with a hyphen. Um uh sour grapes. You know, somebody somebody's didn't get a job that they said they wanted and then they said, 'Oh, it probably wasn't wouldn't have been interesting to live in, you know, in Edmonton, Alberta.'”

A 'mayor of Bloomington' said 'diligent' + 'vigilant' = 'vigiligent' when discussing bike paths and village development, creating a beautiful blend that captures both diligence, vigilance, and a subtle village component; this shows that word blends can be informative and meaningful rather than purely erroneous.

“the mayor of Bloomington was on the radio and he talked He was talking about how how it we have to be vigilant in uh in uh in uh putting in bike paths and other kinds of things in the town. And I thought that was a wonderful blend because it's it's it's vigilant and it's diligent. And it also has a little bit of village in it.”

Phrase blends include examples like 'in one shape' (in one piece + shape), 'easy-go-lucky' (easy-going + happy-go-lucky), 'count my lucky blessings' (count my lucky stars + count my blessings + thank my lucky stars), and 'hit the stack' (hit the ceiling + blow his stack + hit the sack), revealing competition at the phrase level.

“phrase blends. Uh phrase blends. Uh I won't read these out loud, but I think uh well, maybe I These are funny...You know, I hope the package got there in one shape...from a native speaker...He was a real easy-go-lucky guy. Easy going and happy-go-lucky...I should count my lucky stars. I said it on the phone and then I said, 'What? Is that right?' Something Count my lucky stars. Count my blessings, thank my lucky stars...I typed in an email when I was very tired and I wanted to hit the sack. And I said, 'I'm worried that my editor is going to hit the stack.' And I thought that was a marvelous blend of hit the ceiling, blow his stack, and my own feeling that I wanted to hit the sack.”

The phrase 'hope the package got there in one shape' (instead of 'in one piece' or 'intact') was produced by a native English speaker in a typed email and reveals competition between two idiomatic expressions, demonstrating that blending occurs even in formal written communication.

“I have it still. You know, I hope the package got there in one shape. Okay. Uh from a native speaker. Okay.”

Word choice involves competition between multiple candidate words in 'semantic space'; when one word is overwhelmingly activated (like 'dog' when seeing a dog), no error occurs; but when multiple words are nearly equally activated, audible competition manifests as speech errors, hesitations, and word blends.

“the x-axis in these bar graphs is sort of um it's it's a one-dimensional reduction of what you might call semantic space...here I'm facing a situation which might evoke, you know, a word like dog. And so dog is the is the big winner. It's very activated. And and maybe uh fox is a little bit activated...cat is a little bit activated...dog is so it towers above the others enough that it's a big winner and there's no visible or audible...competition. But a lot of the time when we speak there is audible competition.”

A dean at Indiana University said 'will pull no stops unturned' instead of 'will pull out all the stops' or 'will leave no stone unturned'—a complex phrase blend showing simultaneous activation of multiple idiomatic expressions.

“my dean at one time at Indiana University was trying to lure a person to be in cognitive science, and he made this marvelous, marvelous statement that will pull no stops unturned to get him. I mean, this This is amazing.”

A snow shadow is the absence of snow on the ground where a tree blocks falling snow, analogous to and a generalization of a regular shadow cast by an object blocking light; both are instances of the same abstract concept of an obstruction preventing something from reaching a surface.

“there is no sun at all. It's a cloudy day. What is that pat pattern that dark thing under? That is a snow shadow if you wish. That is the absence of snow...The snow has fallen but it didn't go through the tree...it got caught by the tree and so no snow was on the ground and that dark patch is the absence of snow. And it's a kind of a generalization of a shadow.”

Hofstadter's friends Kelly and Dick called his home study an 'office' because in their household, the second-floor room with books, computers, and work materials was called a 'home office', which mapped perfectly to Hofstadter's room; Hofstadter called it his 'study' because his father's prototype had been called a study, not an office.

“my friends Kelly and Dick were visiting my house and um uh they kept on referring to my office. And I kept on thinking, 'What does my office have to do with my house? It's over, you know, across campus.' And then I realized, of course, they were talking about my study...Why do you always call it my office when I call it my study?'...they said, 'I don't know. It's your seems like your office.'...In our house, on our top floor, where we do our work, where we have our books and our computers, uh that's where our office. Uh we have a home office...you have on your top floor...you have computers, you have books, and you do your work there. And it maps perfectly. So, to them it was totally obvious that this was my office. And to me it was totally obvious that it was my study because I took my dad as my prime as my prototype. He on the upper floor of our house up on Moraga, he had a study...And his office was down on campus.”

Thought is the highest level of abstraction-seeking, defined as putting one's finger on the essence of a situation and then bouncing back and forth between the actual situation and the perceived essence found in one's memories.

“Thought is the highest level of abstraction seeking, I'm sorry, the highest level of abstraction. Putting one's finger on the essence of a situation, and then bouncing back and forth between the actual situation and the essence that one found in one's memories. In one's memory.”

The standard proportional analogy format 'A is to B as C is to D' (e.g., 'shoe is to foot as glove is to hand') is commonly used on IQ tests and SATs but represents a narrow, overly formal definition of analogy that misses the true nature of how analogies function in cognition; restricting analogy to this formula puts analogy 'into a little tiny box.'

“analogy is to thinking as a motor is to a car. A is to B as C is to D. And this sort of reduces it down to the old standard, you know, A is to B as C is to D, proportional analogy looks like a fraction... uh shoe is to foot as glove is to hand... these kinds of things that maybe they put on IQ tests and SATs... but that's not how I think of analogy, either. I mean, that's That's That's That's again putting it into a little tiny box and saying that all of analogy is A is to B as C is to D. And it And it again it misses the point.”

In 2005, Hofstadter studied Einstein's analogical reasoning and discovered that Einstein created the concept of the photon by recognizing a mathematical analogy: the black-body radiation spectrum and the ideal gas velocity distribution both have bell-shaped curves, leading Einstein to hypothesize that light, like gas, might be made of particles.

“I spent uh a long time in the year 2005 working on in a talk called Well, it wasn't called, but it was about Einstein's analogy making...what I was really astounded was I had never known how Einstein created the concept of that we now know as the photon...we begin with something that was known this is the basically what's called the black body spectrum...there's different amounts of different energies...people understood...the situation of thanks to Maxwell and Boltzmann of an ideal gas...people knew...the behavior of an ideal gas was due to molecules bouncing around...they could make some mathematical calculations and figure out what the curve should be...a bell-shaped curve...for all the different kinetic energies...at low speeds there weren't very many, high speeds there weren't very many, but at some intermediate speed there were a lot...Now, not this was around for anybody to see, you have to understand, but nobody thought about it. Nobody connected these two ideas until Einstein. And Einstein said this looks similar to me...maybe it's possible these things this thing over here the ideal gas we think as as being made of of corpuscles, particles, molecules. And so maybe it's possible that light is also made of particles...He did a lot of elegant mathematical calculation to to check this out...Basically, a bell-shaped curve and a bell-shaped curve, the analogy between those two...Einstein decided 100 years before me that I that light might be made out of particles. And and that was the what nowadays we call the light quantum hypothesis.”

During a year at a research institute in Trento, Italy, Hofstadter faced the problem of how to greet people of different social relationships in Italian; he learned that 'ciao' is appropriate for close friends, 'salve' for intermediate relationships, and 'buongiorno' for formal relationships, requiring on-the-spot categorization of relationship status.

“I spent a year there...the question that I had to face...was what do you say to these people when you run into them in the corridor? And you don't even know quite who they are...You don't say ciao to people that you are not on familiar terms with...ciao is something that you can say to somebody that you say tu tu, but you don't say ciao to, you know, to the director of the institute...You say buongiorno, or you say salve. And salve is sort of intermediary...every time at every moment as I was walking down the corridors, I had to be doing these on-the-spot categorizations. Is this a a ciao person, a salve person, or a buongiorno person?”

While flying from California to Indiana, Hofstadter read about the Atlantic Conveyor ship avoiding Hurricane Helene by deviating south; upon walking through Denver airport, he saw a woman with a suitcase crossing the corridor and spontaneously made the analogy: 'I'm the boat. She's the hurricane or the whirly girl. This corridor is the Atlantic Ocean.'

“I was reading a book...about Voyage to America...I was in an airplane...he's traveling across the Atlantic Ocean...they discovered a hurricane...they went to the south, they deviated a little bit, and missed it...I get out at Denver. And uh I'm walking down the corridor...here's a picture of me and this is this is a woman with a suitcase...I was swerved a little bit to the left and all of a sudden bang I said, 'Hey, I'm the boat. She's the hurricane or if you wish the whirly girl.' And and this corridor is the Atlantic Ocean.”

The concept of 'chair' has extreme variability; one can purchase books containing thousands of different kinds of chairs, and the variety is mind-boggling, making feature-based definitions of category membership inadequate for explaining how we recognize a chair.

“Chair, I remember there was an exhibit of chairs at the San Francisco International Airport some years ago. Wonderful exhibit of crazy chairs. And then I got books of chairs. You can buy books of thousands of different kinds of chairs and the variety is is mind-boggling.”

Hofstadter as an 8-year-old was fascinated by exponents and subscripts; upon seeing his physicist father's papers with notation like 'X sub 1, X sub 2', he hypothesized that subscripting might be an operation on numbers analogous to superscripting's exponentiation, but was disappointed when his father explained subscripts were merely a naming convention for variables.

“when I was a kid, I was 8 years old as I recall when I found out about uh exponents. And I was in love with math...I made tables of not integers to various powers...I saw a paper uh that my dad was reading, and he was a physicist, and it had math notation galore...there were uh there there little things that said, you know, like X sub 1, X sub 2, and stuff like that...My guess was, 'What in the world What operation on numbers does this subscripting thing do if the superscripting thing, you know, puts it to a power?'...My dad told me, he said, 'Oh, that doesn't do anything. It's just the way of naming variables.' And I was so so disappointed, completely completely flattened.”