Part 2 of a series of (slightly polished/revised) notes I wrote (mostly) while working on a blog post called “What Do You Mean by God?” These notes surround a family of fallacies I broadly call “the denotation switch.” Go here for an index.
5. Some words are (unavoidably) more vague than others. Sometimes vagueness matters.
It’s often clear what a word denotes. If I say “the chair I’m currently sitting on is uncomfortable,” there’s usually no question of what I mean to pick out with the word “chair.” If I say “many chairs have four legs,” we know what type of generic object I’m referring to (of which we might call the chair I’m sitting on a “token”).
If you tell me you don’t know what I mean by “chair” in ordinary uses, I’ll say you’re pretending (I call such moves putting on the philosopher’s hat).
There’s leeway, however, for extraordinary uses, such as when we’re asked in an epistemology or metaphysics course to provide a definition of “chair” that picks out all and only objects that are chairs, while simultaneously satisfying folk-intuitions about what chairs are. A perhaps impossible task.
Often, of course, what a word denotes may be vague even in ordinary usage, and we may argue about what a word really refers to or should refer to, as in the cases of “art” and “pornography” (e.g., when something presented as art strikes many as really being pornography, and government funds are involved). This may be so despite our usually having no trouble using such words in daily life, where we tend to operate in the middle, to speak, of concepts, rather than at the edges. (I can’t define “music,” but I’m certain I’m listening to some as I type this.)
Sometimes, as with “art” and “pornography” in certain contexts, denotative or semantic vagueness matters a lot. Words like “race,” “justice,” “viability” (in discussions about abortion), “democracy,” “free will,” and “volition” come to mind. Politically charged cases can be especially tough to parse out, often due to the difficulty of finding a point of convergence among diverse moral intuitions about what such words pick out either in actuality, or should be understood to pick out (e.g., given some set of socio-political goals).
[This comes up with respect to race in my post, “What Is Race?: Four Philosophical Views by Glasgow, Haslanger, Jeffers, Spencer.”]
Such discussions sometimes clearly are about what a given word should denote. Sometimes the result of this will be the recommendation or official proclamation of an officially or unofficially designated ruling body (e.g., a group of lawmakers, or people who determine what counts as a symptom of a given illness, or an activist organization, etc.).
The point is that sometimes we can’t let terms go loosely defined, no matter how vague. But things often don’t go smoothly.
Activities considered child abuse today would not have been considered so in the 1950s. A sudden expansion of what counts as child abuse can give the false impression of an alarming spike in behaviors that have been commonplace for years—e.g., allowing a child to ride a bicycle without a helmet or to wait unattended in a car, or see this 2015 article by Adam O. Goldstein, MD, MPH: “Is Exposure to Secondhand Smoke Child Abuse? Yes.”
While we might have great reasons to newly designate some behaviors as abusive or harmful, the apparent sudden rise in abuse cases is false. For a great discussion of this phenomenon, but with different specific examples, see Chapter 5—”Kind-Making: The Case of Child Abuse”—of Ian Hacking’s 1999 book The Social Construction of What?
This occurs with mental illness as well. The American Psychiatric Association’s change in the features and diagnostic criteria of a given mental illness may result, virtually overnight, in a much larger or smaller group of legitimately diagnosable people.
Similarly (often overlapping with the above examples), lawmakers are frequently obliged to adopt some operational definition or another of some word, and whatever they land on will be influenced by individual intuitions and interests.
Philosophers, as a group not generally obliged to have a consensus about a word, might debate, say, what we mean—or should mean—by a word. For example, “knowledge,” which generally refers to belief elevated to a special status (e.g., one that vindicates the believer). Should we develop a definition of “knowledge” according to good logical practices and work being done in the cognitive and social sciences, or should we develop a definition according to how people usually mean the word in everyday usage? Is everyday usage at all consistent?
(Not to mention that “belief” itself can be characterized in significantly different ways; e.g., in terms of behavior or experience—the former may allow for computers and thermostats and other non-consciousness things to have beliefs, while the latter does not.)
Scientists must devise operational definitions of the things they measure. They often adapt everyday words when doing so. For a psychological study to measure, say, pity, the concept must be operationalized so that it can be measured numerically. This is tricky business. Maybe impossible business, but it can be useful provided everyone understands what is really meant by “pity” in the study. My concern, however, is that public discussions of such studies generally proceed with no mention of how the concept in question was operationalized.
An actual example this involves the so-called “hot sauce paradigm” formally introduced in 1999, in which aggression is measured against how much hot sauce is allocated to others by unwitting test subjects who’ve just, say, played a violent video game.
[See, for starters: “A Hot New Way to Measure Aggression: Hot Sauce Allocation” by Joel D. Lieberman, Sheldon Solomon, Jeff Greenberg, and Holly A. McGregor; and “The Hot Sauce Paradigm” (brief overview at Improbable Research); and at The Guardian: “Is there any evidence of a link between violent video games and murder?“]
I’m not a psychologist, but often drop in on conversations between psychologists (i.e., on podcasts), and what I often hear are debates over whether a research program’s operationalizing of a vague concept (pity, pride, happiness, violence, jealousy, moral indignation, disgust, racial bias, and on and on) is useful or misleading.
But the specialized use of everyday terms doesn’t have to involve numerical measurement. It’s common practice for scientists in general.
Richard Dawkins is careful, especially in later editions, throughout The Selfish Gene (1976/2016) to emphasize that he is not using terms like “selfish” in the ordinary sense, but as a useful way to talk about the phenomena under discussion. That’s great. But this care isn’t always taken.
For example, I occasionally run into “altruism”‘s everyday meaning used interchangeably with the biologist’s meaning of the word—e.g., characterizing the sting of a honeybee as a hive-defending suicide amounting to an act of kindness, as though the brave bee knew that to sting is to die. Or, here’s a Guardian article that explores whether animals can be altruistic: “Rationing ravens and merciful monkeys: can animals be altruistic?”
This is only a real question if we mean “altruism” in the human-applicable sense. Given the biologist’s definition, however, it seems less difficult a question. As the Stanford Encyclopedia of Philosophy (SEP; accessed 7/14/20) puts it:
This biological notion of altruism is not identical to the everyday concept. In everyday parlance, an action would only be called ‘altruistic’ if it was done with the conscious intention of helping another. But in the biological sense there is no such requirement. Indeed, some of the most interesting examples of biological altruism are found among creatures that are (presumably) not capable of conscious thought at all, e.g. insects. For the biologist, it is the consequences of an action for reproductive fitness that determine whether the action counts as altruistic, not the intentions, if any, with which the action is performed.
That said, my (non-expert) understanding is that this is not so clearcut—or at least has not always been so clearcut—among scientists working in the biological disciplines. For example, there is the (fascinating and tragic) story of George Price, who did indeed seem to mean by “altruism” something human-like (and not only in applications to humans, about whom we can also legitimately ask whether altruism is possible in the “ordinary” sense of the word), as explored in Oren Harman’s 2010biography of Price: The Price of Altruism (I haven’t read it, but it’s on my wish list).
But, in general, biologists do need something rather straightforwardly definable if they’re going to have conversations with one another about goings-on in nature. That’s fine.
But it seems some scientist and science communicators get a lot of mileage out of using words in their field’s technical sense, but without letting people know that this is what they’re doing. Physicists seem to be especially prone to this. I began to notice this several years ago, when I first began studying philosophy, and, after some frustration, came to the cynical conclusion that this was often being done in order to blow people’s minds with what were in reality fairly un-extraordinary ideas.
I’ve recently been vindicated on this count by hearing rogue mathematician Eric Weinstein put this in almost exactly the same way. Right down, I believe, to my hunch that the amount this is done is in inverse proportion to the amount of clout physicists have lost since their rock star days, which are now decades in the past and haven’t been helped by the String Theory program.
When physicists use the word “world” in phrases like “many-worlds interpretation of quantum mechanics” or “branching worlds,” what they’re really using is a metaphor that illustrates for laypersons (and that gives themselves an interesting-sounding shorthand) what comes down to an interpretation of some math (namely, Schrodinger’s equation), or so I’m given to understand as a non-physicist.
I can’t blame them for using interesting-sounding shorthands for their own discussions, but the unqualified use of these terms in public conversations frustrates me.
Many-worlds proponent Sean Carroll often makes a point of justifying his use of the word “world,” as in his 2019 book Something Deeply Hidden. But, as it’s put at the SEP entry “Many-Worlds Interpretation of Quantum Mechanics,” “The concept of a ‘world’ in the MWI … is not a rigorously defined mathematical entity, but a term defined by us (sentient beings) in describing our experience.”
Or see this 2019 Forbes article by Chad Orzel: “Many Worlds, But Too Much Metaphor.”
[Read about my own grappling with the sorts of things I’ve heard physicists say about the many-worlds interpretation, some of which I’m nearly certain they don’t actually believe, but that sure sound mind-blowing when taken at face value: “Some Naive Questions About Many-Worlds Quantum Mechanics.”]
“World” is just one example from physics. Another word to consider is “time,” but I won’t do that here. Same for “information,” as in phrases like “information is fundamental,” which of course isn’t only uttered by physicists.
[Interestingly, I’ve heard Carroll point out that “information” is a “useful handle” we could get along without in a good description of reality. This endeared him to me! On YouTube: “Sean Carroll – Is Information the Foundation of Reality?” (9/1/18).]
I’m afraid the phenomenon comes up in mathematics as well. I won’t explore this today, but at some point in these notes I’ll address “infinity,” and whether, when we look closely at what mathematicians mean when they say that there are different sizes of infinity (e.g., “some infinities are bigger than others”), this really has anything to do with the ordinary understanding of the word.
For example, more carefully, we’d say that there are infinite sets of differing “cardinality” (e.g., those that are countable versus those that are uncountable). And this idea can be expressed in yet more technically careful, further broken-down set-theoretic language.
I’ll save that for its own post, but if you’d like a teaser, consider the rejection of infinity by physicists, which I touch on in this post: “Agustín Rayo Argues for Zero (mathematical platonism vs. nominalism).” Namely, see the discussion of what it means when physicists say that 1 + 2+ 3 + 4 + … (and on and on forever) = –1/12. (This doesn’t address the set-theoretic interpretation, but is roughly in the same genre.)
I’ll close this section by noting the care Bertrand Russell takes in his 1919 book, Introduction to Mathematical Philosophy, to point out, with respect to the Peano’s axioms that, “It is assumed [in Peano’s system] that we know what is meant by ‘0,’ and that we shall not suppose that this symbol means 100 or Cleopatra’s Needle or any of the other things that it might mean” (p 11), where the point is to motivate the idea that, as great an achievement as Peano’s five axioms are, it is necessary to advance beyond them:
…we want our numbers to be such as can be used for counting common objects, and this requires that our numbers should have a definite meaning, not merely that they should have certain formal properties. This definite meaning is defined by the logical theory of arithmetic. (p 12)
In other words, Russell’s efforts (and he is not alone) nicely exemplify attempts to smoothen out, rather than exploit, vagueness, including for the ambitious sort of layperson who might read this introductory book. Furthermore, such a project may be what it takes to learn, after all, the incredibly valuable lesson that vagueness is not so easily done away with.
(Russell also includes three chapters whose titles feature the word “infinity,” a word denoting multiple concepts that he takes characteristic care to keep straight.)
In subsequent posts, I’ll look more closely at specific words. I’ll also spend more time thinking about the denotation switch phenomenon in itself.
The present point is simply that semantic vagueness is a common phenomenon that matters, and that this fact often goes unaddressed, even in situations in which efforts have been made to account for vagueness.
To read more about semantic vagueness, here’s a reading list from a University of Chicago, Depart of Linguistics seminar whose aim is to “investigate the phenomenon of vagueness, with particular focus on the linguistic (semantic and pragmatic) questions that it gives rise to, but also with attention to the philosophical questions that have made it a topic of interest to philosophers for millennia”: http://semantics.uchicago.edu/kennedy/classes/s06/vagueness.html.
I do not, by the way, approach the denotation switch as an expert in linguistics, language, or in the vast majority of the areas in which the words I consider are used. Rather, I approach it as someone who desperately relies daily on words for making sense of the world, while often being confused, or at least vaguely uncertain, of what speakers and writers (expert or otherwise) are talking about.
6. Conditions for Blame/Praise, Punishment/Reward
Suppose the vast majority of us agree that anyone who satisfies a set of conditions C merits having applied to them the label R. Being labeled R may be understood as meriting praise or condemnation, or it may be a morally neutral designation. Let’s suppose here that we’re talking about R-labels that merit harsh punishment.
Replace R with whatever you think worthy of harsh punishment.
Applying R to a person amounts to shorthand for saying that the person satisfies conditions C, where C may not always be clearcut, but is clear enough that we think, for example, that someone who can be shown (beyond a reasonable doubt, etc.) to satisfy C merits, say, at least 20 years in prison.
(Substitute an appropriate punishment, given your chosen substitution for R.)
Now, suppose conditions C are changed to C’, and then C”, and so on, until they are quite different than C. What these variations have in common is that they are, or have been, the conditions required in order for a person to merit the label R.
[Were I being more careful, I’d establish conditions Ci, where i stands in as a indicator of which variation on the set of conditions is being discussed. C0 would be the starting set of conditions, followed by subsequent “generations” C1, C2, and so on. I won’t be so careful here.]
It seems to me that, as C is altered to C’ and so on, we must also reevaluate our widespread assumptions about what constitutes an appropriate response to meriting label R. For instance, in the special case I alluded to above, 20 years may be unambiguously inappropriate. In fact, C might hold in one domain (e.g., the legal domain), and C’ and C” might hold in other domains (e.g., the public or professional domains, respectively). Each domain may, then, merit a different response to R.
That said, it is in fact generally the case that as C is altered, we do indeed generally change the “official” penalties for R along with those alterations, including across relevant domains.
This presents at least one gravely serious worry. Namely, even as our assumed appropriate responses to R are changed in response to changes in C, earlier stigmas surrounding R persist. For example, those who today bear the label R, bear the same label as those who, fifty years earlier, bore that label with the understanding (then and now) that they merited certain harsh penalties (e.g., 20 years in prison instead of the newly revised punishment of an appropriately worded apology).
In other words, altering the conditions for R (rather than assigning an entirely new label, say S), is an efficient way to smuggle in earlier stigmas surrounding R.
[I’ve taken up careful use of the phrase “to smuggle in” after reading that usage in Gary Drescher’s 2006 book, Good and Real: Demystifying Paradoxes from Physics to Ethics. As in: “There need not be any deliberate deception or dishonesty involved in the sleight-of-hand conflation of an unspoken, implicit definition with an explicitly proposed definition” (Kindle Location 268).]
This smuggling in need not be done on purpose or strategically for it to be efficient and harmful, but often does seem to be done on purpose or strategically. In fact, the point of continuing to use R precisely often clearly seems to be expressly in order to borrow gravity from the understanding of what it means to have satisfied earlier conditions for R (e.g., C’ instead of C”’).
There are other serious worries related to altering C while holding R fixed. For instance, 95% of us might not even realize that C has been changed at all, or if we do, might not trust that the change is in good faith, given the above point about smuggling in stigma. This might dilute what would otherwise be appropriately strong intuitions about R.
Similarly, one might consent to having satisfied conditions C”’, only to be then accused of having admitted to having violated conditions C, only to then find oneself informed that satisfying conditions C”’ is in fact more morally reprehensible than satisfying conditions C, if not as legally actionable, and so on—such claims only carry weight due to residual stigmas held over (or smuggled in) from C.
It’s also worrisome that one might admit to having satisfied conditions C”’, thus meriting the label R within a particular domain, but those outside of the domain only see the label R (thus assuming C, not C”’).
And so on. This is only the beginning of the worries connected to switching the denotation of R (i.e., changing the conditions that R signifies one’s having satisfied).
Finally, the above is what I’d call a de re account, in which the label R is applied to a person as a function of that person having satisfied conditions C (within some specified period of time in the past; note that I haven’t mentioned time limitations here, but this, too, is assumed to be stipulated in the conditions defining C, and may be different for, say, C’). In other words, C is seen here as a thing in itself, the existence of which (with respect to a given person), is denoted by attaching to that person the label R.
A de dicto account is also possible, in which R is explicitly understood as “whatever we say it means, no other conditions required” (where “we” indexes one or more persons with the means to do such a thing). When this is the case, the de dicto operations surrounding R must be made very explicitly known, rather than tacitly framed as a de re designation.
I fear that, in general, de dicto operations are often afoot for such designations, but are sold as—perhaps even start out as—de re accounts in order to get a populace to sign up for them.
When this happens (either dishonestly or unconsciously), the denotation switch is dangerously in full force.
Scariest of all is when a populace does in fact knowingly agree to an explicit de dicto operation.
Enjoy or find this post useful? Please consider pitching in a dollar or three to help me do a better job of populating this website with worthwhile words and music. Let me know what you'd like to see more of while you're at it. Transaction handled by PayPal.