31
$\begingroup$

In learning a new area it is very helpful to have high-level intuitive analogies that keep track of the various parts of an important argument or strategy. Experts have a store of such things, and often the only way to hear about them is to talk with the experts or hear the intuition during their talks.

I am talking about very intuitive analogies with the property that one side of the analogy can be understood almost completely by the layperson, but after the "mapping" to the mathematics on the other side, the analogy outlines the argument or strategy reasonably well, of course missing most key details… . The point is to keep in mind a rough outline for the purposes of navigating the technical literature!

For example, in the theory of finite von Neumann algebras, Sorin Popa and his collaborators have made excellent use of his "deformation/rigidity strategy", which Sorin has described in at least one of his talks using the following physical analogy:

Consider a bucket of dark liquid in which you know there is a hard stone. If you put your hands in the water and swish them around and never feel the stone, then you know the stone must be located where you have not swished your hands.

In this picture, the bucket of liquid including the stone is the finite von Neumann algebra $M$. The "liquid part" of the von Neumann algebra can be "stirred" by a pointwise 2-norm deformation of the identity by normal, unital, trace-preserving completely positive maps relative to some subalgebra $A$ which was not "stirred". The hard stone can be, e.g. a subalgebra $B$ with relative property (T), since trying to pointwise "stir" the unit ball of such an algebra by maps of the above sort is not possible without moving the ball uniformly. If we can deform $M$ around $A$ and we know $M$ has a property (T) subalgebra $B$, we can conclude that $B$ must have been contained in $A$ (up to something like unitary conjugacy).

Anyone in the field can see the lies I've told in the above paragraph, but nevertheless the intuitive picture does a reasonably good job of communicating the parts of the strategy. In fact, you can immediately see the main limitations of the technique by asking what happens if (a) there is no stone in the liquid (or the stone is not large), e.g. as in a free group factor and (b) the whole algebra is a stone, i.e. the factor itself has property (T).

Question: What are your favorite such expert intuitive analogies for important parts of your subject? Please include the analogy and explanation of the "details", as I did above.

As in the example I include above, an analogy presented as an answer should include a reasonably complete explanation of the details on the "technical side". The best answers are those which encode and organize surprisingly many technical details in the intuitive analogy, and are not just intuitive mnemonics for remembering the existence of some theorem or other.

EDIT: The following are some helpful modifications to the question suggested by Aaron Myerowitz.

The question starts out with the claim (which is certainly unsupported): "Experts have a store of high-level intuitive analogies that keep track of the various parts of an important argument or strategy."

Question: Is this claim valid? Is it common for experts to have a store of such analogies? Is this more common in some fields than in others?

To loosen the very strict requirement of metaphor suitable (on one side) to the layperson, we may ask instead:

Question: What are some metaphors used to convey the gist of a topic or technique to people not in one's field?

In the latter question we'd like to focus on topics that are not necessarily part of the tool kit of "most" experts, and want to steer clear of descriptions and analogies used to communicate with fellow experts.

$\endgroup$
17
  • 8
    $\begingroup$ Regarding the vote to close, there is probably something I am not communicating clearly. Wouldn't it be useful to have organizing analogies available for more rapidly learning new areas? This would benefit research by allowing us to more rapidly learn important ideas in other areas we are not familiar with. This is very much about research-level mathematics, as it is about facilitating cross-pollination of ideas to accelerate progress. Please let me know what about this is not appropriate or clear! $\endgroup$
    – Jon Bannon
    Mar 20, 2016 at 15:31
  • 8
    $\begingroup$ With the level of analogy that you have asked for, this question can be mapped into Thurston's question, "Thinking and Explaining", of course, missing some key details :) $\endgroup$ Mar 21, 2016 at 14:59
  • 2
    $\begingroup$ @Amir: This made me smile, and I see your point! This helps me understand the votes to close a bit more, thank you! $\endgroup$
    – Jon Bannon
    Mar 21, 2016 at 15:14
  • 4
    $\begingroup$ @Amir: It may be that this question embeds in Bill Thurston's question, or perhaps that Bill Thurston could have provided a nice answer to this question. I find myself rather missing Bill Thurston right now... $\endgroup$
    – Jon Bannon
    Mar 21, 2016 at 19:17
  • 3
    $\begingroup$ I would be happier if the question were much tighter, exactly because there are a million "hairy ball theorem" or "hear the shape of a drum" type answers possible, whereas I think the question is more interesting if it excludes these by demanding that the examples be "organizing" in a strict sense. $\endgroup$ Mar 27, 2016 at 13:10

4 Answers 4

10
$\begingroup$

The analogy between proofs and games. This analogy is so strong that it can be formalized in two mathematical ways.

In (finite) model theory, it gives Ehrenfeucht–Fraïssé games. Logical truth becomes the existence of a winning strategy. This technique underlies many proofs of undefinability and of logical equivalence of structures.

In the Curry-Howard correspondence, it gives game semantics. There are specific analogies that express some very specific concepts e.g. Wadler's devil bargain illustrating how classical logic can backtrack on its choices.

$\endgroup$
6
$\begingroup$

Larry Guth has a nice paper on Metaphors in systolic geometry which I think fits the bill quite well.

In the context of systolic geometry, one such "organizing analogy" that I found helpful is that between a pair of invariants of a manifold $X$. One is the famous Lusternik-Shnirelman category (LS) of $X$, and the other the systolic category (SY) of $X$. The definition of LS is topological whereas that of SY is geometric (roughly, the greatest length $k$ of a product $sys_1\, sys_2 \ldots sys_k$ of systolic invariants of $X$ that can serve as a universal lower bound for the total volume). Here the existence of certain systolic inequalities has led one to conjecture similar lower bounds for LS which were eventually proved by Dranishnikov--Rudyak, Strom, and others.

$\endgroup$
2
$\begingroup$

Might functors between categories (as typed 'n-categories') be properly deemed 'analogies'? When I first learned of categories, I was inspired (having first read Hesse's Glass Bead Game) to deem Category theory as a possible basis for such a 'game'. The question that now remains for me is how properly to define the notion of 'deformation' for arbitrary functors and how to properly define 'approximate analogies' in terms of category theory (if, in fact, such notions can properly be defined in Category theory at all....)

$\endgroup$
16
  • $\begingroup$ This doesn't seem to at all be the sense in which Jon is using the word "analogy" $\endgroup$
    – Yemon Choi
    Mar 29, 2016 at 22:56
  • $\begingroup$ @YemonChoi: The question for me is, whether Jon's example of the "deformation/rigidity strategy" can be correctly expressed in terms of Category theory (I think it might be able to correctly be so expressed). If so, then I might have (in some sense) captured the sense in which Jon is using the word "analogy". In fact, each of the three previous examples can probably also expressed in Category-theoretic terms as well, so Category theory can be used as a language in which analogies can properly be expressed. In what sense do you believe Jon is using the word "analogy"? $\endgroup$ Mar 30, 2016 at 0:03
  • 3
    $\begingroup$ Although the discussion of categories is interesting, and the observation is a good one (I do believe that mathematical structure as expressed in categories precises the concept of mathematical analogy), what I am looking for are "structure preserving maps" from mathematics to the concrete world of human sensory experience, as an exploration of certain ways of stimulating native human thinking for organizing abstract techniques. The "cheese" examples I'm looking for are not a dime-a-dozen, as they should not be contrived and yet should encode and organize a lot of the technical structure $\endgroup$
    – Jon Bannon
    Mar 30, 2016 at 11:40
  • 3
    $\begingroup$ of a mathematical theory. Roughly, what I'm looking for are maps to common sense experience that capture the basic structure of a strategy or technique with a large kernel, namely the definition-baggage of the abstract theory. Admittedly, this is probably mostly for mnemotechnic purposes. $\endgroup$
    – Jon Bannon
    Mar 30, 2016 at 11:41
  • 1
    $\begingroup$ @Yemon: I will consider commenting on the other answers, but I'm at a bit of a loss regarding how to do this in a way that won't transmogrify this thread into a discussion. For brevity, I'll keep quiet for a while and hope people read the comments associated with this answer. $\endgroup$
    – Jon Bannon
    Mar 30, 2016 at 19:30
1
$\begingroup$

Here is an example: Sullivan's analogy (or dictionary) between complex dynamics and Kleinian groups.

$\endgroup$
2
  • $\begingroup$ (But I am not sure if this answers the question as intended.) $\endgroup$
    – Gil Kalai
    Mar 25, 2016 at 9:38
  • 10
    $\begingroup$ Could this be fleshed out a bit; it seems very terse. $\endgroup$
    – user9072
    Mar 27, 2016 at 15:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.