CONTENTS

1   INTRODUCTION

2   EXPOSITION

3   THE HIERARCHY OF REFERENCE

4   THE NEUTRALITY OF LANGUAGE

5   CONCLUSIONS

4.         THE NEUTRALITY OF LANGUAGE

  Having looked in some detail at the assertions A and B of Section 2.1, we now move on to assertion C, namely that language, as characterised by Frege, does not distort or restrict its subject matter. To do this we split the assertion into two parts. The first, based on the word “distort”, looks at the neutrality of language. The second, based on the word “restrict”, looks at its completeness, its ability to handle anything reality could possibly throw at it.

4.1       The “Cake” Problem

 How can an ontology, especially a realist ontology, be based upon a theory of language? Surely language is a human invention, constructed for human purposes of thought and communication. A realist must say that the world is prior to language, and must impose itself on language if that language is to be genuinely referential. If the forms of language are found to be imposing a structure on the world, thereby restricting what we can say about the world, then the claims of realism have been falsified. It might indeed be right to start ontology from the human perspective of the forms of language, but the consequence of doing so is logical relativism rather than realism. What counts as true or false becomes relativised to the particular linguistic and conceptual framework one chooses.

Attempting to build a realist ontology on top of Frege’s theory of reference seems therefore to be trying to have one’s cake and eat it too. It aspires to the certainty of the logic and yet the substance of ontology. These two can be brought together only at the cost of infecting ontology with the structures of language, an infection which threatens to be fatal to the claims of realism.

It could be the case that the Fregean meta-ontology is an inevitable consequence of the objectively necessary structure of language, but that it imposes substantive constraints on ontology. Assertion C, the statement of ontological neutrality, is logically independent of Assertions A and B. If it were not true then there would be a whole range of ways in which the world could be which we would be unable to think or speak about. Analysing the situation more deeply, we can distinguish four different ways in which ontological neutrality could fail.

The last two of these possibilities do not threaten realism. They are not really concerned with the neutrality of language, but rather with its completeness. Nothing in realism rules out the possibility that there could be features or parts of the world which are decoupled from us and hence epistemologically inaccessible to us. This could happen by a physical decoupling. Neutrinos are almost, but not quite, like this. Possibility (c) here suggests there could be a “logical decoupling”, of things from us - things which are invisible to us because they are filtered out by our “linguistic spectacles”. In Possibility (d) the things are not decoupled. We know about them perfectly well, but are reduced to a state of helpless incoherence when we try to express what we know in words. In Section 4.3 below I shall argue that Possibility (d) is a reality. One of the consequences is the refutation of the "because" in Possibility (c). Inexpressibility is not in itself sufficient to make a feature of the world inaccessible to us. The phenomenon can still crash into our senses and make itself known to us, regardless of the inability of language to express what has happened.

4.2       Neutrality and Realism

The threats to the ontological neutrality of language, and thence to the objectivity of the statements expressed in language, come therefore from Possibilities (a) and (b). But before arguing for the absolute neutrality of language, we should note that relative to previous ontologies, Fregean ontology comes a great deal closer to neutrality. Earlier attempts took for granted that there were certain necessary ontological categories, such as space, time and causation. With Frege these are relegated firmly to the status of the contingent furniture of the world. Within his general framework we can contemplate worlds with time and worlds without time, worlds with space and worlds without space. The old-fashioned categories were a serious hindrance to our understanding of the world, because in fact not only were they not necessary, they were not even true. It took the genius of Einstein to lead us into an understanding of what the world was trying to tell us through its phenomena, by challenging and then reversing some of our fundamental assumptions about space and time.

In place of the old categories, Fregean ontology proposes a new set: object, concept, function, relation, truth value, truth function and so on. The great advance entailed in adopting this new set is that, in place of the old ragbag, we now have a clear structure and a clear rationale. Each of the new Fregean categories has a clearly defined place in the mechanism of reference. Each is related firmly to language, rather than being seen as, in some way necessary, features of the world.

The objection remains however: even if the Fregean categories are primarily linguistic, are they not imposed on the world as necessary features by our use of language in speaking about the world? Does this imposition not compromise realism? In a realist account, language should not be prior to the world, but should accurately mirror the world. Language should be driven by the world, and not the other way around.

The need for a prior mechanism of saying arises from the fact that we do not know in advance what the world is going to be like, what our language is going to be called upon to say about the world. We need a prior mechanism which is flexible enough to respond to all challenges the world throws at us. It is precisely the need for a faithful mirror which drives us towards setting up something as general as we can make it, rather than letting the form be dominated by the first facts that come along. (Of course in historical reality, both of individuals and linguistic cultures, experience was allowed to influence the shape of language. It was only the genius of Frege which gave us a linguistic second birth, cleansed of the sins of the contingent.) The faithful mirror is constructed by paying attention not to the way the world is, but to the laws of mirroring (that is, the laws of reference).

The metaphor of the mirror can be extended by comparing a painting and a mirror. Both are two dimensional images of things in three dimensional space. In creating a painting the painter imposes an image on the surface, and must think carefully about the way things are disposed in the depicted world. By contrast the job of the mirror maker is to prepare a surface which can reflect an image of whatever comes along. The mirror maker does not think about what might be reflected, but only about the mechanism of specular reflection. So it is with language: what we say is driven by the world, but for this to be the case, how in general we say things must be prepared in advance.

The metaphor of the mirror is not perfect, because it relates to depicting rather than saying. One could say that names, the residual linguistic pictures, mirror the world. Sentences do something else, something more active. They are used in the mysterious “saying about” activity. Being more active, saying has more potential to violate the ideal of neutrality than the more passive process of mirroring. Yet the highly non-trivial claim of Fregean Ontology is that saying can achieve, for parts if not for all of the world, the linguistic analogue of passive reflection, however actively we must strive to reach this goal.

So far this discussion of neutrality has been more about expressing the thesis, using different forms of words to illuminate different aspects of it, than about defending it. Suppose we ask the opposite question: what would count as a disproof of neutrality? How could we find out that some supposed feature of the world was actually an imposition of the nature of language, and that the reality was otherwise. Ex hypothesi we could not say what that other reality was. But we could know about this reality through direct, inarticulate experience, and would then be frustrated as each attempt to express this reality in language leads us to something other than what has been experienced. This is what I claim happens with the phenomenon of temporal becoming . The attempt to express in language our experience of time (the A-theory) lead to McTaggart’s paradox. Describing time within the canons of Fregean language (the B-theory) avoids paradox, but leads to an account at odds with our primary experience. Proponents of the B-theory, I argue, are suffering from a language-induced delusion about the way the world works, and constitute thereby a counter-example to the neutrality thesis.

We can regain neutrality at the expense of completeness. The parts where language seems inadequate to express experience are excised from the scope of language, and warning signs are posted there saying, “this part of reality is unsayable - attempts to use language here will lead to delusions”. The remaining safe fragment of language is ontologically neutral. This move is discussed further below, under the heading of “Completeness”.

Another, separate challenge to pretensions of neutrality comes from multiple, incompatible analyses (see Section 3.5 above). We find that the same experience has two equally good expressions in language which are nonetheless incompatible with one another.  If they were both genuinely compatible with experience but incompatible with each other, then according to realism one would simply be wrong. But we would be unable to tell which analysis was wrong, and those who happened to use the wrong one would be suffering a language-induced delusion about the world. We would know the existence of this delusion without being able to cure it. To block decisively this disproof of neutrality, we would need to be able to show that the situation envisaged is impossible. This takes us beyond the subject matter of the present essay into the parallel realm of Fregean Epistemology.

4.3       Completeness

It may be that some features of the mechanism of saying prevent it being applied to some features of the world. This does not mean however that we can not claim a high degree of ontological neutrality for the forms of language, as demonstrated above. Nor does it mean we must maintain a Tractatan silence on unsayable matters. (This in any case would not be practicable, as these matters include the phenomenon of time, which is central to our experience and hence central to our communications needs.) There are other means of communication than saying. In practice we deal with the unsayable by incorporating as much as can be said within language, and then using depiction or ostension to do the rest. In principle at least, there are no restrictions on what features of the world can be depicted, as we can used the feature to depict itself.

The classic example of an apparent failure of saying is provided by the passage of time. The frustration of being unable to articulate such an immediate feature of our experience is caught in the famous complaint of Augustine.

“Quid est ergo tempus? si nemo ex me quaerat, scio; si quaerenti explicare velim, nescio.”

As such, this expresses only the failure of articulation. McTaggart’s paradox hints that this might be a necessary failure, that there is something in temporal passage which is permanently inaccessible to language because of the way language must be structured.

This form of failure of saying does not in itself threaten realism, as long as we recognise it when it is happening. It is only if we adopt the Procrustean strategy of forcing all phenomena into the framework of saying, whether they fit or not, that the integrity of reference is threatened. The classic example of this is the B-theory of time - treating temporal extent as static, spacelike, in spite of the centrality of temporal passage to our experience of time. Trying to express in language what passage is leads to McTaggart’s paradox, and thence to the conclusion that passage can not be a feature of reality. The hidden assumption in this argument is that all features of the world must be sayable. Our experience of the world suggests prima facie that this is not the case. The error comes from concentrating too much on language and too little on the phenomenology of experience of which the language is supposed to be a mirror. The error is detected when we notice that what has been reflected by language is a distortion of what is immediately present in our experience, and when attempts to correct the distortion lead to paradox. This is extremely frustrating for philosophers, who are professionally committed to striving to pin things down with words. With the phenomenon of temporal passage, we are reduced to an inarticulate ostension, pointing towards features of our experience and hoping that others will be pointed in the right direction and see what it is we intend.

To explain a philosophical perplexity in terms of a shortcoming of the mechanism of saying could be simply a means of covering up the fact that we are not good enough at philosophical analysis to resolve the perplexity. Some future, greater philosopher may find the required answer. However after struggling with a problem for many years (or millennia) and still being faced with a failure of understanding, it makes sense to entertain the alternative hypothesis that understanding is not possible in this case. Our intellectual effort can then be diverted towards the attempt to prove that understanding is impossible. Such a proof would then forestall all future vain struggles to achieve this impossibility. An analogy can be drawn with geometry: after a long battle to prove Euclid’s axiom of parallels from the other axioms, geometers finally chose the opposite strategy - trying to prove that the axiom is independent. The great value of McTaggart’s work was to begin the process of replacing the struggle with perplexity over time with a proof of impossibility - although this is not how he interpreted his result. More generally the notion of the incompleteness of language opens the way for this type of resolution of perplexity.

Far from being a defeat, the putative incompleteness of language demonstrated by McTaggart’s paradox is good news for realism. It shows that our view of the world is not contaminated by the shortcomings of the mechanism of saying. The realities of the world, in this case temporal phenomena, force themselves upon us regardless of these shortcomings. We then invent pragmatic linguistic patches to enable us to talk about the phenomena in a way which is perfectly adequate for most practical purposes. Only when we go into reflective, philosophical mode do we encounter Augustinian perplexity or McTaggart’s paradox. The conclusion drawn by McTaggart from his paradox, that time must be unreal, is never plausible, because the raw givenness of the phenomena is out there whenever we stop reflecting and open our eyes. This turns us back to look for a problem with language rather than with the world.

5.         CONCLUSIONS

The aims of this essay, with regard to Fregean Ontology have been:

Two sorts of work then follow: strengthening, modifying or even completely replacing these foundations, and building accounts of the world upon them. If, as I believe, these principles make the subsequent task easier and more illuminating, then, I would argue, this is an argument that the principles were correct in the first place. It is not a rigorous proof, but it adds further plausibility to the postulates of Fregean Ontology. This sort of pragmatic utility is certainly not the same thing as truth, but it can, I believe, be counted as evidence for this truth.

Click here to go to the start of the essay