“Flawed” Analogies

Note: This post contains plagiarism: I’ve stolen some content from an earlier blog post, and from my comments on another. I beg the forgiveness of faithful and diligent readers.

Recently I’ve had to deal with some complaints from people on Twitter who seem to have misinterpreted certain analogies. Worse than that, sometimes it seems as though they don’t understand why and how we use analogies at all. Here are some examples.

I suggested, “Waiters don’t tell the chef when the food is ready to serve; why should testers tell the managers when the product is ready to ship.”

One fellow objected, claiming that waiters only deliver the food, and that to use that analogy diminishes the role of the tester. He implied that testers would be more like food-tasters who would inform the chef about the quality of the dish and possible problems with it. Fair enough—but even then, the chef would have responsibility for the decision to deliver the food.

Another fellow objected, demanding “Do you mean that testers aren’t first-class members of your dev team?” Well, as a matter of fact they are—but now I have some questions about whether this person might not see waiters as first-class citizens of a restaurant. As Milton said, “They also serve who only stand and wait.” (Yes, that’s a joke.)

Someone else objected when I said that “Testers are responsible for quality in the same way that investigative reporters are responsible for Supreme Court decisions.” The response was, “This surprises me. Don’t we try to get testers involved up front, w/lotsa communication & collaboration, to improve quality?” Yes, we do. The point here is simply that, whatever their activities, responsibilities, or contributions, testers are not responsible for making decisions about quality, but rather for informing decisions about quality. Like investigative reporters might inform the evidence presented in a court case, or in issues of social policy. That’s still a completely valuable role in a society, but it’s not the decision-making role that we would associate with those who make the laws or provide enforceable interpretations of them.

Analogy, metaphor and simile are kinds of associative language. Let’s start with “analogy”. The Oxford Dictionary of English (available as a killer app for the iPhone) calls analogy “a comparison between one thing and another, typically for the purpose of explanation or clarification…; a correspondence or partial similarity (my emphasis there)…; a thing that is comparable to something else in significant respects”.

The word comes from the Greek. “Logos” refers to words, thoughts, and reasoning. The prefix “ana-” denotes several possibilities, including “up, upward, throughout, backward, back, again, anew”. Analogy, therefore, suggests thinking something through, or reasoning back to some similarity, or looking at a thought again, or thinking something up. In other words.

For metaphor and simile, I’ll repeat here a quote from an earlier blog post: In his book The Educated Imagination (based on the Massey Lectures, a set of broadcasts he did for the Canadian Broadcasting Corporation in 1963), Northrop Frye said, “Outside literature, the main motive for writing is to describe this world. But literature itself uses language in a way which associates our minds with it. As soon as you use associative language, you begin using figures of speech. If you say, “this talk is dry and dull”, you’re using figures associating it with bread and breadknives. There are main kinds of association, analogy and identity, two things are like each other and two things that are each other (my emphasis –MB). One produces a figure of speech called the simile. The other produces a figure called metaphor.”

If it helps, think of an instance of associative language as a kind of model. Models are heuristic representations—literally, re-presentations—of something complex in terms of something simpler; of something unfamiliar in terms something more familiar; of something more abstract in terms of something more concrete. “Heuristic” is a key adjective here, suggesting that models help us to learn, and are fallible. As Nassim Nicholas Taleb says in The Black Swan, “Models and constructions, these intellectual maps of reality, are not wrong; they’re wrong only in some specific applications.” Absolutely true—and you can flip it around to say that models are right only in some specific applications too. It is their very imperfection—the fact that they do not represent the thing being modeled in every dimension—that makes models useful and powerful. The relative simplicity of the model is an aid to our understanding of a more complex subject. The purpose of models and associative speech is not to create an exact parallel, but to get people thinking in terms of similarities and differences. Were there a one-to-one correspondence between every aspect of a model and the thing being modeled, there would be no simplification and no point in using the model. It would be like Steven Wright’s full scale map of the United States: “It took me all last summer to fold it.”

Metaphor and simile trigger the capacity of the human mind to forge new links between ideas. Frye again: “The poet, however, uses these two crude, primitive, archaic forms of thought in the most uninhibited way, because his job is not to describe nature but to show you a world completely absorbed and possessed by the human mind… The motive for metaphor, according to Wallace Stevens, is a desire to associate, and finally to identify, the human mind with what goes on outside it, because the only genuine joy you can have is in those rare moments when you feel that although we may know in part, as Paul says, we are also a part of what we know.”

If I were to tell you that my dog were like my car in that they both require a lot of maintenance, would you expect my dog to have a gas tank? Would you expect me to take the dog to my mechanic when it appeared ill? If I said that picking up the guitar again after a couple of years was “riding a bike” for me, would you take me literally and envision a musical instrument with wheels and handlebars? If I were to suggest that my former boss was a pain in the ass, would you expect everyone at work to have trouble sitting down? If I said that a friend is as honest as the day is long, would you think that I measured honesty in hours? If I said that your newer testers were green, would you contradict me because they didn’t match a certain Pantone colour?

Analogies are always flawed by design. They’re intended to make people think about the ways in which something is like something else, while intentionally ignoring the ways in which they’re different. That is, they’re not supposed to be perfect. They’re supposed to draw attention to some factor of something, so that people can consider that factor, and possibly others, in a new or different light. Notice that I can use that expression, “in a new or different light”, and you don’t get confused or think “Hmmm…incandescent or fluourescent?”

Complaining that an analogy is flawed is like complaining that a model airplane (say, one designed for a wind tunnel) is flawed because it’s not big enough to carry real passengers. Confusing an analogy with a fact in this way is a symptom of model blindness, evidence of a failure to distinguish between models and reality. That can bite people in two ways: first, by limiting our imagination to the literal, rendering us capable of less powerful reason, and even making us something less than human (which is what Frye warns against); and second, by allowing us to place too much trust in our models, making us vulnerable to risk when then model doesn’t apply (which is what Taleb warns against).

Some people say that they don’t like analogy, or that they don’t use analogy. To me, that’s a symptom that they’re not paying attention to the way they speak and the way they think. If you’re such a person, I have a suggestion: try paying attention to the number of times per day you describe or envision something in terms of something else. Try monitoring how often you try to represent an idea with a sketch. Note each time you say, “That’s like…” Get a friend to observe each time you use a metaphor. You’ll shortly observe that we’re all using analogy, all the time. Then try quitting for a day, like fasting or giving up cigarettes for Lent. (Did you notice? Those were similes. Did you understand them and absorb them, or did you say, “I don’t smoke”?) Not only will your speech or writing be limited, but I would argue that your capacity to reason will be limited as well.

If you’re in software development, the program that you’re writing or testing is a way of modelling some task that the user wishes to perform. The ways we have of describing those programs are universally framed in terms of models, analogies, metaphor, and simile. We humans are analogy machines (Did you notice? That’s a metaphor.) So if you don’t like “flawed” analogies, it would nonetheless be a good idea to get used to them.

Postscript, 2013/12/10: “A study published in January in PLOS ONE examined how reading different metaphors—’crime is a virus’ and ‘crime is a beast’—affected participants’ reasoning when choosing solutions to a city’s crime problem…. (Researcher Paul) Thibodeau recommends giving more thought to the metaphors you use and hear, especially when the stakes are high. ‘Ask in what ways does this metaphor seem apt and in what ways does this metaphor mislead,’ he says. Our decisions may become sounder as a result.” Excerpted from Salon.

Another postscript, same day: have a look at Surfaces and Essences: Analogy as the Fuel and Fire of Thinking, by Douglas Hofstadter and Emmanuel Sander for a look at how pervasive and fundamental associative speech is.

9 replies to ““Flawed” Analogies”

  1. “So if you don’t like “flawed” analogies, it would nonetheless be a good idea to get used to them.”

    Sure. But *sometimes* when I say that an analogy is flawed, it doesn’t mean that I’m confused about the use of analogies themselves. And it doesn’t mean that I don’t use analogies a lot.

    Instead, it *might* mean that the specific analogy in question was a poor one.

    If you decide to say “testing is like a warm puppy”, you might get general agreement, or some might express that it was a poor analogy.

    Michael replies: I’m not sure that it’s helpful to say simply that it’s a poor analogy. Poor compared to what? By what standard? By whose lights? So let me propose an alternative: rather than saying it’s a poor analogy, I might say, “I’m not clear on the parallel you’re attempting to draw here,” or “I’m not sure what factors that you see in common between this thing and that thing,” If you look at the middle and the bottom of this post, you’ll see some tests that one might apply to a metaphor.

    Analogy is a form of communication. Sometimes it is done well, sometimes not so well. Just the fact that it is an analogy/metaphor/simile doesn’t imply successful communication.

    That’s right, it doesn’t. But you might be thinking of analogy as the end of something. I prefer to think of it as the beginning. Even a terrible analogy can get us thinking and talking, and that’s the object of the game for me. The model is the trigger, not the silver bullet.

  2. Aren’t all user interfaces (including command line) similes? They express what the computer is doing, or what we want it to do, but we’re not directly interfacing with the hardware – there’s at least one abstraction layer between us and the electronics. Thus, any software tester is exposed to simile, whether they like it or not?

    Also, this reminds me of an episode of Star Trek, where the crew encounters a species who communicates using only metaphors. All previous encounters had resulted in conflict, in part because the context of the metaphors didn’t make sense.

    Context is important when dealing with analogies.

  3. I use analogy (mainly humour-based) a lot – and occasionally write about it, here. For me, it’s about tackling communication from different perspectives (almost a focus-defocusing technique).

    I find Monty Python a great source for thinking about misinterpretation, fads, fanactical devotion cults and holy test cases, an example as applied to software testing being here.

    Agree completely about the danger of getting trapped by models – and I find analogy is a way (for me) to occasionally break free of the shackles.

  4. I tend to use analogy when trying to understand what the system I’m trying to test is supposed to be doing. I think it is easier for analogies you create for your own understanding to be interpreted correctly by others; as opposed to an analogy you are using to explain a concept to someone else. As with your Chef and Waiter example above, if you came up with it for your own understanding, you are aware of the limitations you are placing on the modal. Others are not aware of this limitation, and so take the analogy too far, thus thinking you have said something you haven’t. The question then is, how do you stop an argument brewing about the use of the analogy, instead of keeping to the discussion that the analogy was brought in for in the first place?
    The other problem that stems from taking it too far is that some people hear “X is like Y” instead of “X is like Y for this specific (i.e. limited) reason”. As an example, if I said, “Testers are like spiders, in that they create a web of tests to catch bugs”, some people will ignore the limit of the analogy and think I said, “Testers are like spiders”, and then apply other ways in which they are alike. They could conclude I was saying people have an irrational fear of testers (although, in some cases that may be valid :-)), or that tester are annoying little critters. I’m not trying to say that, but those are examples whereby the analogy can be taken too far (and possibly too literally).

  5. I think that it’s much more to do with the low self-esteem of your readers and not about analogies. They are trying to coerce you into conformity.

    I am surprised you bit. 😉

  6. Note: This comment starts with plagiarism from another blog.

    Analogies are powerful when they help us understand something (they shouldn’t be used to argue.)

    Great post, but it’s perfectly fine to argue against an analogy/metaphor, especially if a discussion renders more insight.
    You can state your reasoning, as in this example:

    Michael, I disliked your last comment “We humans are analogy machines”.
    Sure, it highlighted the fact that we use analogies a lot, but it uses the word “machine”, which I believe is an unfortunate analogy in software business.
    Since our industry create new things all the time, I think factories and machines should be avoided in our language as much as possible.

  7. I think metaphor and analogy are good for communicating tacit knowledge. However when writing we attempt to convert the tacit to the explicit. Some may pick up on what you are trying to communicate if it resonates with their understanding, and others may pick up a different dimension communicated by your “tasitness” . This can be a good thing. Also some people are more “S” than “N” and don’t communicate the same with respect to metaphor and analogy – people don’t all pick up the extra information in the same way. After your metaphor leaves your bow it is no longer under your influence.


Leave a Comment