The objectivity of meaning
13 May 2005
That meaning can be explained in terms of rule-following (the rules that govern our use of words) is not something we have too much trouble to accept. The idea that "meaning is use" can be understood in a wide variety of ways. But if we take it to imply something along those lines, then it does not appear to be so controversial after all. In particular, it seems that such a view suggests an account of meaning in which this notion retains its essential aspects of normativity and objectivity. Rules are obviously normative and it does not seem outlandish to think of them as objective too. All it takes is an account of how a certain behavior may be correct or incorrect according to a certain rule "independent of what I or anyone may think of it"(1).
What comes as a surprise, though, is that these premises, albeit being apparently Wittgenstein's own views about meaning, seem to lead straight into the rule-following paradox, and its meaning evaporation consequences(2). Explaining why it does has not proved easy. What we can gather from Wittgenstein's text is that this stems from a misunderstanding; namely to think that there is no way of grasping a rule which is not an interpretation. According to Kripke, and those who acknowledge his influence, this reply amounts to a "sceptical solution" to a "sceptical problem". Yet, others maintain that there is no problem at all, provided we get rid of our habit to view "signs as dead" and picture them as "alive" instead(3). Does all this makes things any clearer ? I do not believe it does. My impression is that the prolific debate which followed the publication of Kripke's seminal text did not result in any real clarification of the matter but instead gave rise to a variety of competing positions each of which is rather less illuminating than the original text of Wittgenstein.
The claim I wish to make here is that the order-processing example, and other such examples drawn from MIS practice, provide conceptual means to genuinely clarify the issue of rule-following and even, in some ways, to achieve a clearer understanding of it than Wittgenstein himself could. In the section titled "the grammar of 'in accord with the rules'", we saw that (quite surprisingly) everyone agreed that the system was not in accord with the rules (Qa) but that there was no agreement on whether it was in accord with the specification (Qb). In the previous section, we said that this discrepancy may be explained in the following manner: Whenever we attempt to compare a particular course of action with a symbolic human artefact, be it a piece of text, a picture, a wooden model, a string of words learned by heart, there is the potential for the kind of infinite regress which leads to the paradox (in the same sense as "whenever we transship large quantities of gasoline, there is the potential for a fire"). What we just called a "symbolic human artifact" is, in my view, what Wittgenstein has in mind when he uses the phrase "an interpretation of a rule". Henceforth, we will just call it a specification(4).
What the rule-following paradox shows is that a specification is not a rule; i.e. that if we allow a specification, at some point in an argument, to take the place of "the rules", we are in danger of witnessing the dissolution of the whole concept of "following a rule". The consequences for meaning is this: the meaning of words does not resides in specifications. If we allow our view of meaning to include specifications as a direct source of the meaning of words then the whole notion evaporates.
In a strict sense, "semantic Platonism" is not directly threatened by the rule-following paradox. Positing the existence of abstract entities corresponding to words meanings does not result in the paradox by itself. Platonism is a problem only inasmuch as, once abstract meaning entities are assumed, it is almost impossible to resist the temptation to create artifacts in their supposed image which one can then use as a specification. It is quite obvious that Platonism was not created as a response to the rule following paradox. However, it may still appear now as a last-ditch measure to stop the infinite regress by conferring taboo status on a particular specification one is desperately unwilling to abandon. Taken in that sense, Platonism or "classical semantic realism" is indeed directly at odds with the paradox.
Other meaning theories, which do not involve abstract entities, are threatened by the paradox in a far more direct manner. One of them is Logical Positivism. What are the verifiability conditions that were supposed to allow the assignment of a truth value to a statement if not a specification of the method to follow for that purpose? Indeed, Logical Positivism may be viewed as the purest form of specification-based meaning theory. The Logische Aufbau der Welt is the ur-specification or, at least, the ultimate toolkit to build perfect specifications. Nothing is left in Logical Positivism except specifications. It should come as no surprise that so little remains of it.
As we said at the end of the previous section, even Kripke's presentation of the supposed "solution" to the paradox in terms of assertability conditions, runs the risk of leading into the same trap. Once it is assumed that "all that is needed to legitimize assertions that someone means something is that there be roughly specifiable circumstances under which they are legitimately assertable", the step taken in the writing down of an actual specification of those conditions, does not seems to be very big. But if such a thing is done the dreaded "sceptic" will appear sooner or later to challenge one word or another used in the document which will have been thus produced. And we are led back to our original perplexity.
My goal is not to make a complete list of all the meaning theories supposedly threatened by the rule-following paradox (though I think it is indeed quite long). The point I wish to make is that their very number is a testimony to the power of the impulse that leads us to make the confusion between a rule and a specification.
How do signs come to have meaning?. This is, no doubt, a valid question and one that deserves to be investigated. Once this question is taken as a stating point it may seem that a logical second step is to rephrase it as such: "how can a sign that is meaningful determine its applications?". And the same may be asked about rules ("how can a rule determine its applications?") or expressions ("how can a meaningful expression determine its applications?"(5)). Independent of why these reformulations may be viewed as warranted or natural, it should be clear by now that they are dangerous and should be avoided. While there is a certain sense in which a rule determines its applications, the same cannot be said of a "meaningful expression". An expression is a definite string of signs, i.e. an instance of what we previously called a specification. And we have seen that it is in attempts to judge courses of action (or "applications") against such artifacts that lies the source of the paradox. Using two sentences of the same form, one with "rule" and the other with "meaningful expression" in its place is an invitation to the same kind of confusion we saw occurring with (Qa) and (Qb)("is the system in accord with the rules?" vs. "is the system in accord with the specification?"). If we take it for granted that an answer to the question "how do signs come to have meaning?" must take the form of an investigation into how it is that a "meaningful expression" can determine its applications, then it should not come as a surprise that we find it difficult to avoid the paradox. If we are looking for specifications we will find specifications; and, with them, the kind of infinite regress that takes place whenever we try to judge acts directly against them.
Signs are dead. A quick, unbiased, look at the marks on this page is sufficient, I believe, to convince oneself of this. The only thing we can possibly mean when we attribute meaning to them is that they are used in the context of a certain practice in which a specific role is given to them. We use the word "rule" sometimes to refer to a certain definite subset of such practices; like e.g. in "a new rule has just been added to the accounting standards for bonds". But, quite often, we simply use the phrase "the rules" to refer in a very loose and general way to all the practices which are relevant in a given context. It is in that sense, for instance, that the system of the order-processing example is in violation of "the rules". That the meaning of this phrase is completely devoid of any idea of a specification in this case is made apparent by the fact that the correct way to process internal invoices is nowhere said to have been determined by any single document or "rule expression". The procedure for internal invoices presumably emerged through a number of remarks made in meetings, e-mails, memos, as well as specific actions taken at certain junctures. We may, for example imagine the newly appointed CEO of the newly acquired subsidiary barging into the office of the chief accountant of the parent company and yelling "Hey! you just invoiced us at the regular rates. How are we supposed to make money in these conditions?". And then the answer might have been "Oh yes, X told me about that but my assistant forgot to follow the new instructions". Or it may be that this problem had been completely overlooked and that this exchange was therefore the very event that triggered the setting up of the new procedure. Most people may have forgotten these events, or never known about them. Nonetheless, everyone knows that there is a rule for internal invoices and what is the correct way to follow it. It is a commonplace to say that the most resilient and reliable rules are generally unwritten. So why are we still tempted to confuses rules with specifications?
There is one possible use of the word "rule" which seems not to have been considered by Wittgenstein in the Philosophical Investigations. Imagine a superior saying to a subordinate:
(O) "henceforth, regarding issue XYZ, the rule will be such and such".
This cannot possibly refer to customs in the sense of the established practices of a community. Yet it is clearly a valid use of "rule". And it is still possible to make sense of it in terms of practices, if not in terms of customs. It refers to the future acts the superior commits himself to: praise or reward if the rule is followed; reminders, blame, disciplinary action, and so on, otherwise. This is an instance of a clearly non-objective rule. At first, it rests solely with the superior to decide whether any action undertaken by the subordinate is in accord with it. Many will no doubt protest at such an endorsement of arbitrariness and counter that the objective aspect of meaning precisely guarantees that anyone could perform the same judging function or even that, in fact, the correctness of the subordinate's behavior according to (O) is independent of what anyone thinks about it. That this view is mistaken is a direct consequence of the rule-following paradox. What the superior has done so far is merely to utter a string of words. He is therefore still free to interpret them in more or less the way he pleases. The customs which apply to him will play a role in determining whether the margin of interpretation at his disposal is wide or narrow. If this margin was nil, it would mean that the interpretative regress could not take place, and there would be no paradox.
However, as time passes, the superior comes to have less and less leeway to interpret his initial statement. People have witnessed him make repeated judgments regarding a large number of cases. One may imagine cases in which this practice has turned out to be quite different from what he himself originally intended. But as the number of cases and people involved increases, his practice becomes increasingly stable and predictable. Later on, one may even imagine him leaving his post and be replaced by someone else. Then, at some point, all his subordinates may also have left and been replaced. None of the original people involved in the creation of the rule remain but the rule itself is still there; probably quite well entrenched and undisputable by now. Meaning rules are like this one, but even to a greater extent. They have been around in a sufficient expanse of time and space to be well beyond the control of any individual in particular, however powerful he may be. In that sense, they are objective. They have become, from anybody's point of view, as rigid and unchangeable as objects. Or even more so. It is arguably easier to remove mount Everest from the surface of the earth than to remove "Bachelors are unmarried men" from the rules of the English language.
There is therefore finally a sense in which meaning can be said to be objective. But it appears to be only relatively so. The wider the community, and the more well ordered it is, the more objective meaning is within it. But this notion of objectivity is certainly not the one we were looking for initially. In particular, according to it, correctness is not independent of what anyone thinks. It is, however, independent of the opinion of anyone in particular. Even if no present member of a community considers a certain act as correct, it is always possible to count on some (or even a majority, or even all) members of the same community to be of a contrary opinion in the future.
Should we continue to search for an absolutely objective account of meaning? Maybe. But in doing so, we should be careful not to advocate theories which have a counter productive effect on our own practices, as it is the case, in my view, of most of today's MIS development methodologies. Language rules and customs in general may not be absolutely objective. But they are very strongly so, as the remark on mount Everest shows. They also have the capacity to evolve, as we have seen by examining the case of statement (O) above. By contrast, the rule following paradox shows that a certain category of attempts to formulate absolutely objective rules results in a situation where anything actually becomes possible: "And so there would be neither accord nor conflict here". Zero normativity, of course, but zero objectivity too.
Notes
1. C. Verheggen: "Wittgenstein's rule-following paradox and the objectivity of meaning", Philosophical Investigations 26:4 (2003), Blackwell. This section, is in large part, inspired by this paper and may be considered as a reply to it.
2. "Wittgenstein does not come up with the claim that language is a practice in order to dissolve a problem about meaning. That claim is what creates the problem to begin with". C. Verheggen, Op. Cit. Not everyone would agree with this claim, of course. Nonetheless, I think it is useful to to put it that way. First of all, things appear in that order in Wittgenstein's Philosophical Investigations: first "meaning is use" then the rule following paradox. Above all it seems quite natural: first "meaning as a set of rules for word use" then "but what does it mean to follow a rule?" and, finally, the paradox.