[I am just thinking out loud here; apologies if at times I am non-precise, or fail to use, or even incorrectly use, some element of technical jargon relevent to this subject.]
Sometimes, it is tempting to think, it is possible to disagree with everything a person says, and yet to be completely in agreement with them.
Consider two questions: first, whether it is murder for an individual to kill a person who would themselves otherwise be sure to unknowingly kill them (imagine, for example, someone about to unwittingly engage the machinary that will crush a person to death); second, whether all murderers should be condemned to prison for life.
These appear to be two very different, albeit tangentially connected, questions.
Imagine, however, that Alice answers the first question ‘no’ (that is, she believes that killing in self-defence is not murder, even if the victim did not intend to kill), and the second question ‘yes’ (all murderers should get life-sentences). Bob, meanwhile, answers the first question ‘yes’ (it is still murder even when done in self-defence), and the second question ‘no’ (some murderers should not get life-sentences).
It is possible in this situation, then, that Alice and Bob in fact agree on what should be done in all cases. In non-self-defence killings (without other countervailing factors), both agree on life-sentences for the perpetrators; both agree that there should not be life-sentences where the killing is purely in self-defence. Alice and Bob therefore agree on everything in practical terms, but their agreement is conceiled by the argument between them over the theoretical questions.
Meta-theoretically (so to speak), this situation results from the combination of two dilemmas. The first is an example of what, following W, we may call a ‘conflict doubt’: the question “should we call it murder when it is done out of self-defence?” may puzzle us because some of our rules about the word ‘murder’ point one way (murder as killing without legal authority, or as killing the innocent), while other rules point the other way (murder as malicious, murder as chosen (and choice not applying where the alternative option is death)). Some people may feel no doubt for this precise question, but the question could probably be adjusted to find their own point of definitional equipoise (suppose the victim, for instance, intends to harm the soon-to-be-killer, but does not realise the harm would be fatal; or suppose the soon-to-be-killer reasonably believes the soon-to-be-victim is trying to kill them, though in fact they are not). In any case, situations of societal doubt, or conversational doubt, where the protagonists in a disagreement cannot come to an easy consensus on the matter, can arise even when an individual is able to resolve the question for themselves. Such doubts, W says, resolve into indecision: they are debates in which we – individually or collectively – must decide how to use a certain word. The second dilemma of the pair is then a practical dilemma that makes use of the term about which there has been doubt in the first dilemma. The significance of the second dilemma therefore depends on which path has been chosen in the first dilemma.
Such issues are conversationally problematic, because it may appear that the two parties are disagreeing over fundamental questions of ethics, when in fact their practical beliefs appear to be the same. [Note that the conflict doubt does not inherently have ethical consequences. Bob and Alice may agree entirely on the morality of the case in question, but simply disagree whether the word ‘murder’ is an appropriate description of it]. In particular, if Alice and Bob do not explicitly raise the first question, the question of the definition of murder, they may not understand that their passionate disagreement over the second question masks concordance in the realm of practical beliefs.
And now we can imagine a third person, Clive. Clive agrees with Bob over the first question, the question of murder – they agree that this form of killing in self-defence is murder. But Clive agrees with Alice over the second question – they agree that all murderers should be given life-sentences.
On the face of it, Clive is a moderate, who holds a compromise position somewhere between Alice and Bob – he agrees with part of what Alice says, and with part of what Bob says. And if the two questions are raised independently – if Bob only confronts the first question, and Alice only confronts the second question, each may believe they have found an ally in Clive, and will prefer Clive to Alice/Bob. If Clive is a politician, for instance, who succeeds in focusing Alice’s attention on one part of his platform and Bob’s on another, he may be succesful in gaining their votes.
But of course, in reality Clive’s practical positions are directly opposite to those of Alice and Bob. Regarding this man in the newspapers they have been reading about, who had to kill a man to save himself from death at the hands of an industrial dough-mixer, not realising that his victim was not himself actively trying to kill him… well, Alice and Bob agree that he should not be handed a life-sentence, but Clive believes that he should. So Clive may on the surface, and truthfully, give the same answers to some questions as does Alice, and to some other questions as does Bob, but in practice his positions are exactly the opposite of theirs.
This may be innocuous, where Alice and Bob (and Clive) are able to quickly see that the issue at hand is how these two questions work together – how their beliefs about the punishment for murder depend on how they define murder. But in many cases, they are not able to see this at first. The definitional dispute that gives rise to the apparent practical disagreement between Alice and Bob, and the apparent partial agreement between Clive and Alice and between Clive and Bob, may be, as it were, several stops down the road, hidden behind several layers of intervening beliefs and arguments. Indeed, in many disagreements people likely do not begin with clear and obvious ‘definitional’ questions, but rather the definitions in question lie hidden within the interactions between other beliefs. The ‘conflict doubt’, then, may never in fact be opened up into open conflict, but remain a concealed ambiguity, shaping the significance of all related questions, without anybody ever addressing it explicitly. In these cases, it may be extremely hard to see when a superficial disagreement about one thing in fact masks an underlying agreement on practical terms, warped by an unstated disagreement over the extent and limitations of this or that word. This is particularly the case in real arguments between two independent positions, in which many words may be used in subtly different ways, and as a result the practical ramifications of the positions as complete systems of judgement may be extremely hard to unravel for the external observer. Sometimes, indeed, what appears to one participant to be a dilemma of definition is taken by another person as a practical decision, while what the first person believed was a consequence of a definition is taken by the second person to itself be a definition, and a source of consequences. Or perhaps some people are not entirely clear, even with themselves, which of their beliefs are definitions and which are consequences. Perhaps some are both.
And here is another practical consequence of this. When Alice and Bob argue on one question, Alice may apply great pressure to Bob to conform to her point of view, or at least to appear to conform – if we imagine Alice is in a position of power, either personally or through others who share her views. If Bob concedes the point, Alice may feel victorious, and virtuous – she has persuaded someone with noxious views to see the error of their ways and accept her, more admirable, point of view.
This feeling is understandable. But if Alice has indeed converted Bob on one issue alone, then she has not created a fundamental agreement between them. Quite the contrary. By persuading Bob on one superficial point, she has not made him agree with her in practical matters – she has made him agree with Clive! Where there was an underlying consensus between Bob and Alice, of which they were not aware, her ‘success’ in converting him to her point of view on one question has in fact created a real disagreement between them on practical matters. A disagreement of which they will not be aware unless they continue to engage with one another on other matters, or discuss practical cases.
We may imagine that Bob has not actually been converted, but merely concedes the point in public. But now his public position is that of Clive, and he will be (unintentionally) adding to the pressure that people feel to agree with Clive. Every person of Bob’s persuasion, with whom she has no underlying disagreement but only a difference in arbitrary definitions, whom Alice persuades to publically concede her narrow point, becomes a voice furthering the overall position of Clive, with whom Alice has a real disagreement. Of course, in practice these discussions will not be so straightforward, and easy to analyse. Likely there are some real disagreements between Alice and Bob all along; they are merely less significant than those they have with Clive. The point remains, however, that by attempting to enforce orthodoxy on how one particular question is answered, Alice is not in fact bringing people to agree with her in practical matters, but rather the exact contrary of this. And unless the participants in this discussion freely discuss their answers to a wide range of questions – and are allowed to do so for each question, even when some answers appear ‘unacceptable’ to the consensus, due to fundamental differences in meaning, the counter-productive effect of this process of orthodoxy-imposition will never be uncovered.
Of course, it’s not only murder where these issues can arise. Consider one of W’s better-known philosophical connundrums: is a tomato a fruit or a vegetable?
We can pair this with a second question: should minced beef and garlic and onions ever be served in chopped-up, cooked-down fruit?
Alice, we may suppose, believes that a tomato is a vegetable. She finds the idea of cooking beef and garlic in fruit astonishing, and viscerally revolting – she is thinking here of course of beef mince in boiled apples, or perhaps some slurry of reduced strawberries. She is horrified when Bob suggests cooking such a meal – Bob appears to have no problem at all with cooking beef and garlic and onions in chopped fruit. But this is because Bob believes that tomatoes are a fruit – albeit a rather odd sort of fruit in some ways, taste-wise – and because he likes his rudimentary attempt at spaghetti bolognese. Bob believes he has found a friend in Clive, meanwhile, because Clive agrees that beef in fruit in this way is perfectly delicious. But Clive really is thinking of strawberries – Clive isn’t thinking of tomatoes, because Clive agrees with Alice that tomatoes are not a fruit. Clive agrees with Alice on one question, and with Bob on another, but really it is Clive who is fundamentally at odds with the pair on the practical question of what to eat for dinner. And if Alice persuades Bob that he must never cook meat in fruit, she will never again have that delicious beef ragu that Bob used to make for her.
We may distinguish then between what we might call categorical opinions – opinions expressed in terms of, and helping to define, certain terms, certain categories – and particular opinions – opinions regarding what to do in a particular case. Should John be given a life-term? Should we have ragu for dinner? These are particular opinions… and we may agree on these even if many of our categorical opinions – the opinions that, at least in questions of ethics and justice, we often think of as the foundation and justification of our particular opinions – differ wildly. Conversely, of course, two sets of categorical opinions, alike in almost all regards, may yield wildly differing particular opinions, if there is only a small but crucial difference in fundamental definitions.
And then perhaps we might want to note something odd: we cannot explain the connection between our particular opinions without resorting to the use of categorical opinions. This is why I chose the word ‘particular’. Any time we attempt to explain, ‘oh, I always do things like this in situations like that‘, we must always assume certain categorical definitions when we try to find words to specify this and that.
We may want to have a third category – general opinions – to express our belief that our particular opinions are not invented case-by-case but rather form coherent patterns. We would like to say it is not merely that we want to put this murderer in jail and that murderer in jail and so on for every murderer we are told about – we would like to say that we have a general opinion that murderers should go to jail, as a general rule. But we cannot explain or describe these general opinions except in terms of categorical opinions. In many, perhaps most cases, we would like to say that our general opinion is in fact the result of our categorical opinions. Perhaps we might say that this is the difference between something chosen – like the belief that murders should go to jail – and something unchosen, like a desire to eat beetroot. The general opinion about murderers flows from our categorical opinions about murder, justice, punishment, protection of society and so on, whereas our general opinion about beetroot – the summation of our particular opinion that eating that beetroot would be good, every time we see beetroot – does not seem to flow from categorical opinions, but only to be a persistent habit, a continuity not through entailment but through mere inertia of tastes.
And yet it seems strange to make categorical opinions so central, when as we have seen we may share our general opinions even when our categorical opinions differ through-and-through – indeed, perhaps it is easier to share general opinions when our system of categorical opinions is wholly different, rather than when it is almost shared. Perhaps we might want then to cast off categorical opinions as superfluous – neither necessary nor sufficient for our general opinions. We can believe that John, and Mike, and Ted should all go to jail for life without having the categorical opinion ‘murderers should all go to jail’, and we can also believe ‘murderers should all go to jail’ and yet not believe that John and Mike and Ted should all go to jail, even if there is no controversy over the facts of their cases. Or perhaps again we should merely cease to focus on specific categorical opinions – it is then the categorical system that entails the general opinion. To be sure, the same general opinion may be arrived at non-categorically, through some unconscious instinct, but it seems easy enough to distinguish these two cases in a fundamental way (but is it so easy?). The meaning of the definition only takes its full form – and hence only takes on its practical implication – within the context of the entire system of definitions. [In the case of science, we may speak of Duhem-Quine, and of Kuhnian paradigms]
Then, however, we may wish to examine this concept, complete categorical system. It is hard enough to gather what this may be in a practical situation – people do not preface their discussions with a disclosure of their complete categorical systems! But what can it even be in theory? We may imagine, for example, that a person may only infrequently call to their own attention many of their own definitions. This span between re-assertions of such definitions may be years (even if we assume for the sake of argument that it happens at all for all definitions). So if a person expresses an opinion, within that span of years, what does it mean? We must rely for its meaning on the complete categorical system, including those parts that have not recently been re-affirmed, nor will be in the near future. Can we simply take the ‘last known value’ for all parts of the system? Suppose a person’s opinion is later shown to have been ‘wrong’, whatever we can agree on this meaning. Given the complete categorical system they appear to have held at the time, their statement has a practical significance that appears… to disagree with reality, and to disagree with facts that the person does not deny. Fair enough then, they were wrong. But suppose the person says “oh, that’s not what I meant!” – “But it must be what you meant,” we say, “given your complete categorical system”. “Oh,” they explain, “that’s what I would mean by that if I said it today, to be sure, and what I would have meant by it if I had said it the day of my marriage. But right then, when I did in fact say it, in those odd few years immediately after I was born again, I used the words with quite a different meaning.”
What do we say to him then? Now, if we have him on tape discussion his definitions of things the day of the Objectionable Statement, we may put it to him – perhaps not undeniably, but at least persuasively – that he was being inconsistant. That he did indeed mean the words that way, and hence that his statement was inaccurate. But suppose that we do not. Suppose that the definitions in question are definitions that, quite believably, would not require conscious consideration for long periods at a time, and that the man did not in fact think about how he was defining the terms. How can we say then that he was wrong, and not merely using the words a different way? And not only that he was wrong in the sense of his words, as commonly being understood, being misleading, but that he was also wrong internally, that he was mistaken in his reasoning and not merely in his unusual choice of expression? The difference between his having been right and his having been wrong depends then on the exact content of his ‘complete categorical system’ at that time. But this is not merely a practical problem, of how we could know what was in his head at the time. It is a deeper problem: given that we agree that he was not consciously thinking of the relevant elements of his categorical system at that time, what does it mean, and what is the significance of it, to say ‘the categories he was employing were these, and hence his conclusion was incorrect’, rather than ‘his conclusion was correct, hence the categories he was employing were these‘. Given this problem, I am tempted to say that these two responses seem more like different approaches we may choose, rather than theories only one of which can be correct. And it seems to me that this is not a rare problem, but a continual one. Most of use do not consciously define for ourselves our categorical systems on an ongoing and updated basis; none of us do, perhaps none of us can, consciously define for ourselves our complete categorical system, and yet as we have seen it is only the complete categorical system that gives anything we say its practical significance – any element left unspecified may result in widely different implications.
And then again, all this assumes that disagreement over definitions only arises between individuals, and not within an individual. We may like to say that there are times when we ourselves do experience doubt over whether a word is appropriate to use in a certain case – whether, that is, the definition stretches so far. We might wonder whether such doubt could ever be comprehensively eliminated: for any set of rules about how to use a word, whether reasoning by deduction or by induction, it would seem that a borderline case could be proposed regarding which the rules are ambiguous. If such ambiguity and doubt is allowed to persist in even a small part of our categorical system, our categorical system is not complete, and hence our meaning is not determinate. And indeed it would seem that any attempt to address borderline cases through rules (rather than through the obviously futile task of addressing each individual case one-by-one) would require us to make use of other categories – our rules would themselves have a meaning dependent upon the complete categorical system of which the disputed definition formed a part, forming a vicious circle of ambiguity.
I do not mean to suggest that there are no viable solutions to these problems – perhaps we might insist on shared categorical systems and rule much of what we say meaningless (where our system is not perfectly shared), or perhaps we may simply accept our own ambiguity, and seek instead to find a way to limit that ambiguity to a tolerable level – though this would appear difficult. Or perhaps we may do away with general opinions entailed by categorical opinions and attempt to argue in reverse, from general to categorical. There are no doubt other options also. Yet none of them seem entirely attractive, at first glance.
These ruminations began with a practical problem; I think they have ended in more troubling, theoretical waters, as deep-seated practical problems often do.