Can you have an intuition without an account of meaning?
1.0 Intuitions without accounts
One of the more interesting questions that came out of my discussion with Michael Huemer was this:
Can you have intuitions about whether some statement is true or false independent of an account of what it means?
My discussion with Huemer would suggest that my answer is no. But here, I’d like to give a qualified yes. This question was prompted by remarks here. The issue in that context is whether one can have an intuition about whether something like the following statement is true or false:
It is morally wrong to torture babies for fun.
The answer may be a yes depending on what one is having an intuition about and depending on what an “intuition” is. If the intuition in question is a judgment about what an ordinary speaker would say in response to this statement in an ordinary context, then I think the answer is yes: people would tend to say this statement is true. In the conversation with Huemer, however, it was unclear whether this is what I was being asked, or whether I was being asked something else. My goal in this post is to elaborate on why I take issue with questions like these.
Why don’t I give an unqualified “yes, that statement is true” when asked whether it’s true that it’s morally wrong to torture babies? After all, I think baby torture is morally wrong, and I would say it’s “true that it’s morally wrong” if someone were to ask me on the street. However, what I mean by this reflects my own philosophical commitments. I am not expressing this vis-a-vis what I take to be the semantic commitments implicit in ordinary moral language, such that in saying that the statement is true I am taking a stance on the semantics of ordinary moral claims. I worry that Huemer (and perhaps other philosophers) may not distinguish between proprietary meaning when responding to such questions and ordinary semantics, making it difficult for me to respond to such a question when posed in a philosophical context: I am not sure if the person asking the question is asking me about:
What I would say conditional on my philosophical views and my personal usage of the relevant terms
What I would say insofar as I intend to invoke uses of terms that comport with their use in ordinary language
What I would say conditional on the presumptive interpretation of the relevant terms and concepts as intended by the person presenting the statement
My answers would be (1) true, (2) true, and (3) it depends. In the third case, I’d need more clarification but the answer could be true, false, or I don’t know, it’s indeterminate, etc. Note that (1) and (2) are actually truths about different sorts of claims, so all three responses are technically responses to different questions.
1.1 Contextualized and decontextualized statements
In a philosophical context, it’s reasonable to expect clarification about which of these questions (or some fourth alternative) someone wants me to address. As such, given that the context in which Huemer posed the question was one I took to be ambiguous, how could I have an intuition about whether the statement was true?
Technically speaking, I don’t think statements, independent of some interpretation or other, can be true or false. Perhaps I can isolate a couple operating principles I tentatively employ in these conversations:
There can be no judgments (of truth or falsity, etc.) without interpretation.
And there can be no interpretation without context (real or imputed).
Were someone to use a statement in some context, I’d do my best to judge whether what I took them to mean was true or false. But Huemer wasn’t using the statement. He was effectively mentioning a statement, independent of anyone actually asserting it. In other words, he wasn’t making a statement in context, but referencing a statement outside any particular context of usage. When a statement is presented outside its context of usage, we have something akin to a use/mention distinction. That is, when I am asked to judge the statement:
It is morally wrong to torture babies for fun.
…nobody is asserting the statement. When a person uses a statement, there is some presumptive fact about what that person intends to communicate, and/or facts about what they are doing or how their words are functioning (whether they are aware of it or not). I locate meaning exclusively and entirely in these facts; the statement itself doesn’t mean anything. Rather, the statement is used by that person to mean something. This can be captured in sloganized form as:
Words don’t mean things. People mean things.
So when I am asked whether a statement is true or false, but nobody is asserting it, technically speaking I think this is nonsensical. Statements, when used, are used in context for some purpose or goal on the part of the person presenting the statement. They serve some function. There are no further facts about what the statement itself means, independent of such usages. I view statements as behavioral events or occurrences. Facts about such events are distinctive to those events. There may be statistical regularities or patterns in kinds of events, and this is clearly true with language: two instances in which native English speakers say “It is raining,” are very likely to mean something very similar. But again, these are statistical regularities in common usage. The statement “It is raining” doesn’t mean anything independent of these contexts of usage; it’s just that humans use similar sounds to mean similar things in similar contexts, and we pick up on, learn, and recycle those patterns: indeed, that’s pretty much what languages are; recycled behavioral templates for conveying meaning in (relatively) reliable and recurring ways.
1.2 What can change the nature of a tiger?
My view of language is similar to my view of the analysis of any particular organism’s behavior. Suppose we wanted to know about the disposition of a particular tiger. Let’s call this tiger Tony. Is Tony friendly? Aggressive? Curious? Facts about Tony’s disposition will depend on…well, Tony. What does Tony do? How does Tony act? All facts about Tony’s disposition depend on actual details of that actual tiger’s behavior. Of course, Tony may be a typical tiger, or an unusual one. Maybe most tigers are aggressive, but Tony is friendly. Tigers vary, and there are no fixed facts about how a tiger must behave. Now suppose you were asked:
What is the disposition of a tiger?
How would you respond? If you think of a typical tiger, you may think: aggressive, predatory, cautious, enjoys swimming, sleeps a lot, and so on. But suppose I responded:
No, I am not asking you to imagine the disposition of a typical tiger. I am asking you to judge the disposition of a tiger simpliciter. A tiger in the abstract, as it were.
If this sounds like a nonsensical request, it’s because it is. Abstract tigers don’t have dispositions. They don’t have behavior. Only actual tigers have dispositions. One can attribute an imaginary disposition to a fictional tiger:
Imagine a tiger named Luna. Luna is a very inquisitive and friendly tiger that enjoys scratches and playing.
In other words, you can stipulate whatever features of a tiger you want. But abstract tigers don’t, in virtue of being abstract tigers, have any particular disposition or other characteristics unless these are stipulated (this stipulation may occur in virtue of some stipulated, distinctive concept of a “tiger” or can be appended to a reference to a tiger in the abstract, as is the case with Luna above).
Statements are just like this. There is a difference between a statement as an actual instance of a person uttering something in some context, i.e., some actual event in which Sam says:
It’s morally wrong to torture babies.
…and a statement as a string of unused words nobody is asserting in any particular context. Let’s call the former contextualized statements and the latter decontextualized statements. Contextualized statements are analogous to actual instances of specific tigers. Facts about what any particular contextualized statement means are distinctive to that statement, just as facts about the disposition of a tiger are distinctive to that particular tiger. In contrast, decontextualized statements quite literally cannot mean anything, any more than an abstract notion of a tiger can have a disposition.
Personally, I’d prefer to dispense with the notion that decontextualized statements should even be considered statements. After all, the abstract notion of a tiger isn’t a tiger. So why refer to decontextualized statements as if they were, in fact, statements? They are, at best, abstractions, or statistical composites, that could be used to categorize and allow us to neatly reference actual instances of contextualized statements. But what we’d be doing when discussing what these decontextualized statements “mean” is, at best, drawing conclusions about the statistical trends of any given language communities to use the sounds/letters that resemble these decontextualized statements in particular ways.
In sum: there are facts about what people mean when they make contextualized statements, and there are statistical trends in what people mean when they do so, which could in principle be summarized by and referred to by decontextualized statements. But decontextualized statements quite literally do not and cannot mean anything, because meaning isn’t located in the statements themselves, but in their use.
1.3 Undifferentiated truths
Nevertheless, people can and do discuss decontextualized statements. What should we do when we’re given decontextualized statements? Well, if statements have no meaning outside a context of usage, we must impute meaning onto it. I can draw on my inferences about typical meaning about ordinary language users, by imagining some plausible ordinary context. Or I can draw on what I mean when I use the words in question. Or I could, if presented with the statement, suppose that the person presenting it intends for me to interpret it in the way they understand it. That leaves me with at least three options for how to impute a possible meaning on a decontextualized statement:
My own meaning (that is, some formal account I’d give that may apply in one or more contexts)
The person who presented the statement’s meaning (some formal account they’d give)
An inference about its typical meaning in some ordinary context.
This is what I have in mind by an undifferentiated truth: one cannot meaningfully assign any truth value to the statement in question without some inferred meaning. One doesn’t need to be given an account by the person presenting the statement, but unless the statement is interpreted in accord with one or another possible meanings, it wouldn’t make any sense to judge it to be “true” or “false.” After all, we’re not judging the words in the statement to be true or false, but what we take those words in the statement to mean.
In my discussion with Huemer, the statement wasn’t being used; he wasn’t asserting the statement, but presenting it for my consideration. When this occurs, I believe one has no choice but to impute some interpretation onto the sentence. And if, in the context of a discussion, it’s unclear which interpretation is intended by the speaker, then how am I supposed to know what my interlocutor wants me to respond to in judging the statement to be true or false?
Without clear contextual cues to disambiguate the speaker’s intentions in presenting a decontextualized statement, we run into a problem: there is no distinctive context, so the person evaluating the statement must fill in the contextual vacuum by making inferences about how the person presenting the statement intends for their interlocutor to interpret it. This doesn’t occur in ordinary contexts where a statement is asserted: I will make an inference about what that person means.
If it’s an ordinary person, I’ll make whatever inferences I make about typical meaning in the relevant context. But suppose I’m at an academic conference and a philosopher is making the claim. If I know the philosopher is a realist or antirealist, I may be inclined to interpret their remarks in accord with their philosophical commitments. Finally, if I know that the philosopher holds a stance about the meaning of moral statements in ordinary language, and they ask me to judge whether the statement is true or false, do I interpret the statement in accord with what I believe to be their account of its meaning in ordinary language, in accord with their philosophical commitments (if they differ from ordinary usage), or in line with my own commitments? Simply put: context matters, and asking me to judge an unasserted sentence without any context is a fruitless or at best highly fraught exercise.
2.0 Independent of what account?
There are still more issues with addressing the original question Huemer posed. At least one question is what it means to have an intuition “independent of an account” of what something means. This is ambiguous. Independent of having the person presenting the statement to you provide an account? Independent of imputing your own account onto the statement? The relation between the “account” in question and the statement isn’t specified here. Second, what’s meant by an “account”? I don’t think you need to have a comprehensive theory or analysis to render a judgment about a claim, yet Huemer’s interpretation seems to suggest he may have taken me to be saying something like this. If so, then I don’t think that. What I think is very simple:
In order to judge whether a statement is true or false, you need to understand what’s meant by the statement.
Is that what I’m being asked? If so, there are going to be deeper problems here: while I think there are facts about whether most people would say such a statement is true or false, I don’t think such judgments appeal to some shared set of concepts implicit in ordinary thought and language that would allow us to work out some distinctive, shared semantic core of meaning that the words used in the statement share in common. I don’t think language works that way. So it may very well be that Huemer’s claim that we should be able to have intuitions about whether statements are true or false implicitly relies on a conception of language that I reject. If so, I don’t think I’d be able to have intuitions about such statements.
Another way to put this is that I reject the notion of there being some distinctive semantic facts about the meaning of words and statements. I think it’s pragmatics all the way down; statements don’t have any meaning at all outside their contexts of usage. As such, if I’m being asked to evaluate the truth or falsity of the statement independent of any context of usage, then I regard this as literally impossible, and while other people may report having intuitions about such statements, since I think all of those intuitions would be predicated on misconceptions about language and meaning, all of those people would be making a mistake (even if it were psychologically possible for them to “have intuitions” about these cases). I, at least, would be unable to engage in such a task, since it would require me to adopt presuppositions that I find not simply false but nonsensical.
In short: if Huemer or another philosopher asks me whether I think a normative moral statement like “it is morally wrong to torture babies for fun” is true or false, I don’t know what they are asking me. If that seems absurd, I hope the preceding remarks will partially rectify that impression, but since it probably won’t for some people, I’ll say more.
3.0 Me want cookies!
Why don’t I know what Huemer means? Because Huemer is presenting the statement in a theoretical context, rather than an ordinary context in which familiar contextual cues would prompt me to make inferences about what I’m being asked that draw on my familiarity with ordinary language. Since I think that what statements mean depends on their contexts of usage, if Huemer poses a question in a particular context, any interpretation I have of the statement will depend on my inferences about what Huemer means by the statement in that context.
It is standard practice for philosophers to ask us to judge whether statements are true or false outside any contexts of usage. Our goal is to simply judge whether the statement itself is true or false. It should be abundantly clear by now that on my view this makes no sense. Perhaps if the tiger comparison wasn’t helpful, this comparison will be:
What determines the intent behind an action? The intentions of the person performing the action, or the action itself? I believe it is the intentions of the person performing the action. Attributing intentions to actions themselves makes no sense. Just the same, attributing literal meaning to statements themselves makes no sense. This is barely even a comparison: contextualized statements are actions and are performed intentionally; so why not examine the intent of the people making those statements, and the contexts in which they are making them as an indication of what’s meant by the statement? Why think the meaning is located in the statement, independent of how it’s being used? This strikes me as bordering on magical thinking about language, and the fact that it seems to underlie so much of analytic philosophy is as baffling as it is distressing.
We encounter many contextualized statements in everyday life. People assert things and we have to judge whether what that person means by that statement is true or false. Perhaps a politician says, “I did not embezzle campaign funds,” or perhaps your nephew says “I didn’t eat all the cookies.” These statements occur in contexts and we infer their meaning given those contexts. As you encounter recurring patterns in terms and phrases, you develop a sense of the typical meanings to which people put certain words and phrases. That repertoire of knowledge about typical meaning is built up over numerous encounters: we learn our languages in their natural contexts of usage, and so our inferences about what people mean are rooted in those experiences. This allows us to extrapolate, considering what people would typically mean in typical contexts even if no contextual information is provided. As a result, we can consider decontextualized statements. For instance, we can ask what it would mean to say:
Mark ate all the cookies.
Sandy punched Susy in the face.
It is raining.
I am the walrus, goo goo g' joob.
None of these decontextualized statements are contextualized statements anyone is uttering in any actual context. And yet I am not sitting here, reeling, mystified by these statements, muttering under my breath “What could these statements possibly mean?”
No. Why don’t I have such a problem? Because I can draw on my prior knowledge of the use of these terms to impute a prototypical context in which these statements would be used onto these statements. In most typical contexts, the first statement would mean that there was a person named Mark, there was some specified subset of cookies denoted by “all the cookies,” and that Mark ate all of them (I could elaborate on what I take “ate” to mean, “all” to mean, and so forth, but that would be tedious). We could ostensively gesture at some of the terms here, pointing to examples of things like what we take the terms to refer to. For instance, my first inclination would be to think of chocolate chip cookies; these serve as a kind of “default cookie” or “standard cookie” in my mind:
Whatever kinds of cookies you think of, it’s not reasonable to interpret Mark as having eaten literally every cookie in existence. Why? Because nobody has ever done this before and it isn’t a thing anyone could feasibly do. All the cookies probably wouldn’t be meant in a typical case to mean literally all cookies. As you can see, one’s assumptions about what someone would mean in some ordinary context flesh out how you interpret the statement. And this can be used to circumscribe the use of “all” to some plausibly limited set of cookies: all the cookies at home, all the cookies in the cookie jar, etc. Critically, the statement doesn’t have to mean this. Suppose I gave some more context:
Day 6: It all started nearly a week ago. A silence fell over the world. The birds stopped chirping, animals fled or hid from sight. And then, without warning, a great rift opened in the sky, as a furry blue monstrosity burst through the clouds, its shaggy frame visible from one horizon to the other. It stared at us for a moment before shouting “ME WANT COOKIES.” Nobody knew what to do. The governments of the world attempted to speak with it, but it simply repeated “ME WANT COOKIES.” I think it took a bite out of Arizona. They say half the state is just…gone. The next morning, the governments of the world threw every weapon they had at it. This had no visible effect. It just took a bite out of Russia and repeated “ME WANT COOKIES.”
Day 43: It began with the factories. Nabisco. Famous Amos. Milano. You name it. But they’ve run out. 24/7 production kept it appeased but there’s no more sugar, no more flour. We simply don’t have the resources to keep making cookies. People have started to call the creature “Mark.” I don’t know why.
Day 45: It started up again. “ME WANT COOKIES.”
Day 302: It has been nearly a year since “Mark” arrived. Every government has collapsed. There is no internet, no electricity, nothing. I haven’t found clean water in several days. The Chipgangs are out again, rounding people up and forcing them to make cookies out of whatever we can find: old cans of corn, boxes of crackers and candy. Whatever works. But with the collapse of civilization, we can’t sustain enough cookie production to appease it. I don’t know if this will be my last entry. The Chipgangs may find us. We’re hiding in an old warehouse.
Day 512: That’s it. Mark ate all the cookies. We are doomed.
Now consider what the statement “Mark ate all the cookies” means here. Does it mean all the cookies in the cookie jar? No. Now, with this context, the passage can be interpreted far more literally: Mark ate all available cookies on earth that could reasonably be given to it. Even here, Mark may not have eaten every cookie that exists. Maybe there’s a stale cookie in an abandoned cupboard somewhere on earth. Or perhaps there are other worlds with alien bakers making some form of a cookie.
Why the silly example? Well, why not? Aside from that very good reason, my point is that when we assess decontextualized statements we may be inclined to think the statement itself means something, because it’s just so obvious what that statement means. But that seemingly fixed meaning is an illusion: it is only, at best, provisionally stable, contingent on the specific set of experiences and familiarity with past usages that allow you to impute a typical context on a decontextualized statement. Were you to have radically different experiences, a statement like “Mark ate all the cookies” could change from a cute anecdote about a child over-indulging to an ominous portent of the apocalypse.
Context matters. When a philosopher presents a decontextualized statement in a philosophical context, I react differently than if a nonphilosopher asserts a contextualized statement in some ordinary context. My reaction to the two will naturally differ. Absent disambiguation in the philosophical context, it is not clear what the philosopher is asking. A request for clarification in such contexts is entirely reasonable, since such contexts will often be far more ambiguous than contextualized statements in familiar, ordinary contexts.
4.0 More at stake
There is more at stake in this particular conversation. I have concerns about what Huemer takes intuitions to be. I take the rise of the term “intuition” in philosophy to be an unfortunate trend. Whether one can have an “intuition” about statements without having an “account” of the statement would depend on what’s meant by an “intuition.” There’s a lot of variation and imprecision in how philosophers employ the term: it may be that on some accounts, there are no intuitions of the relevant kind, and so one couldn’t have intuitions of that kind. In other cases, the notion of an intuition may be too unclear to give a definitive answer. In still others, the answer would be a qualified “yes.” Depending on what Huemer means by “intuitions,” I may very well think nobody has any intuitions at all: that the notion of “intuition” in question is a bit of muddled pseudopsychology that doesn’t pick out any real psychological states anyone is actually having.
Or perhaps Huemer means something I’d find acceptable. In that case, I still take issue with the suggestion that one can have intuitions about a statement without having an account of the statement: that itself strikes me as ambiguous. I don’t think the matter is resolved much by where Huemer takes the conversation. Huemer goes on to give an example that purports to show that there are clear instances in which a legitimate philosophical question can be posed about that does not presuppose an account of what’s meant: Gettier problems. Recall the initial question:
Can you have intuitions about whether some statement is true or false independent of an account of what it means?
In some instances, the whole point of a question is to assess how you would use some term or concept in some specific situation. In that case, the purpose of assessing the statement is to determine how you would use a particular term or concept in that particular instance for the purposes of coming to a better understanding about the content of the concept, either on your personal and potentially idiosyncratic use, or insofar as the goal is to identify the features of some putatively shared concept.
In the case of Gettier problems, the whole point of these questions relies on these assumptions:
We have some (possibly uniform, shared) concept of “knowledge”
The concept has specific application conditions that are consistent with the conceptual features that define, or constitute, what knowledge is
These application conditions are opaque; we don’t have direct introspective access to them
We can try to discover what those application conditions are through the method of cases. By considering our intuitive reaction to various edge cases and potential counterexamples, we can determine what our concept of knowledge is
Thus, when presented with a Gettier case, one’s judgments about whether it’s “true” that one has knowledge provide insight into one’s own implicit concept of “knowledge.” As such, can one have an intuition about whether a statement is true or false without someone telling them in advance what knowledge is?
Yes. Absolutely. While I have reservations about the use of the term “intuition,” I see no problem at all with someone judging whether this or that case is an instance of “knowledge,” and you obviously don’t need an account of knowledge to make such judgments. I don’t think that in such cases our intuitions are necessarily tracking a shared concept; they may instead tell us something about ourselves. And I definitely don’t think that such intuitions are tracking some sort of transcendent, immutable, a priori truth about the essence of knowledge itself.
For comparison, I don’t need to know the rules of grammar to judge whether a particular sentence is grammatical or not. My knowledge of grammaticality is implicit: I know it when I see it, even if I can’t articulate what the rules are, and even if I wouldn’t recognize those rules if they were presented to me. The same could be going on for many philosophical judgments, insofar as those judgments are taken to be judgments about our dispositions to use concepts in certain ways.
If this is what Huemer was asking me when asking me whether I think it’s “true” that torturing babies for fun is wrong, then I think I can judge whether the statement is “true” or “false.” However, there’s a big problem here, with respect to using the Gettier problem. In the case of Gettier problems, the whole point of the problems is that we don’t have some shared, explicit analysis of what “knowledge” is, and the purpose of the case is to attempt to suss out what our implicit concept of knowledge is.
Yet when it comes to asking me (someone who works in metaethics) to judge whether it’s true that baby torture is wrong, it’s implausible that I am being asked to engage in the same kind of procedure. After all, is Huemer asking me to engage in the method of cases to determine how I employ the concept of “immoral”?
That would be a pointless exercise, because I already have a post-theoretical commitment about the concept of “immoral.” I’m not engaging in, nor interested in engaging in, the analogous quest of determining what the folk concept of “immoral” is via the method of cases. First, I don’t think this is an appropriate method for figuring out what other people mean when they say things. I take such questions to be empirical questions. If this request presupposes that there is an a priori answer to such questions and that one can figure out that answer via intuition, then I reject that we can do this, and I question what, psychologically speaking, would even be going on with people reporting having “intuitions” about such cases. Such intuitions may be no more truth-tracking or substantive than claims to see auras or hear the voice of God.
Even if we set aside issues with what an intuition might be, it might very well be that Huemer or others are expecting me to have a semantic intuition that presupposes a uniform and determinate, shared semantic core to moral claims. Since I don’t even think semantics of this kind is a thing, if people are using some kind of faculty that inclines to judge claims on the basis of such assumptions, then I think everyone doing this is engaged in massive error.
There may still be a question about whether people could have some kind of psychological response we could call an intuition in response to such cases: perhaps a disposition to judge statements presented outside a context as “true” independent of having any clear sense of what they’re doing or why. I can’t deny that people could very well behave this way. But I’m not sure what to make of what would be going on in such cases. That seems like yet another empirical question.
In any case, Gettier problems are intended to sort out some putative “ordinary notion” of knowledge. When considering questions about the truth of statements about the wrongness of baby torture, I am not engaged in some analogous analysis of the ordinary meaning of good or bad. I take such questions to be empirical questions, just as I take questions about how people use the word knowledge to be empirical questions. In short: the Gettier problem is a paradigmatic instance of using the method of cases to explore the application conditions of an implicit concept. Questions about whether it’s true that baby torture is right or wrong either (a) aren’t, in which case Huemer’s appeal to such cases is a bit strange or (b) are, in which case that isn’t the task I took myself to be engaged in and, if it were the task Huemer wanted me to engage in, relies on presuppositions about language and philosophical methods that I reject (and that I’d also reject for Gettier cases). If all I were being asked, instead, was about my own concept of “right” or “wrong,” rather than some implicit concept, then comparisons to Gettier cases are even more strange, and, at any rate, I have an explicit position so I don’t need to use the method of cases to report what my stance is: I am an antirealist, and when I make moral claims my claims are consistent with my commitment to antirealism.
So if Huemer were intending to prompt me to engage in the same activity as what one is up to when engaged with Gettier problems, he’d be asking me to engage in a kind of intellectual exercise that I think is completely misguided and almost entirely irrelevant to any points I was seeking to make in that discussion, at least conditional on my views about language and meaning being correct. After all, I don’t think that discovering how I use the words “immoral” or “bad” tells me much about how nonphilosophers use those terms. Such cannot be resolved, on my view, by evaluating my intuitive response to decontextualized statements.
Now I’m going to go have a cookie.