Online research in philosophy

We propose a shift in perspective from the view of natural languages as formal languages to natural languages as a collection of resources for constructing local languages for use in particular situations. This is suggested by our experience constructing natural language grammars for particular applications using the Grammatical Framework. It points to a research programme investigating how such resources play a role in linguistic innovation by agents constructing situation-specific local languages and how they can be made dynamic, modified by the linguistic agent’s exposure to innovative linguistic data.

This essay considers what it means to understand natural language and whether a computer running an artificial-intelligence program designed to understand natural language does in fact do so. It is argued that a certain kind of semantics is needed to understand natural language, that this kind of semantics is mere symbol manipulation (i.e., syntax), and that, hence, it is available to AI systems. Recent arguments by Searle and Dretske to the effect that computers cannot understand natural language are discussed, and a prototype natural-language-understanding system is presented as an illustration.

Why is it that philosophy seems unable to obtain the kinds of agreement regularly achieved by mathematics and the natural sciences? The experimental philosophy movement emphasizes conflicting intuitions as a potential source of philosophical disagreement. This essay draws attention to another, complementary source: the logical imperfection of natural languages. Unlike logic as it is formalized in symbolic notation, the rules governing the correct use of terms in a natural language can be indeterminate, underdetermined, and inconsistent. Though most philosophers recognize the logical imperfection of natural languages in the abstract, everyday philosophical discussion is often conducted as though the argumentative moves agreed upon as legitimate, applied to the raw materials of the discussion, should lead toward agreement. Yet the logical imperfection of natural languages makes it possible for disputants to agree upon natural language premises, adhere to the usual, uncontroversial rules for drawing inferences from those premises, and yet arrive at conflicting conclusions. The essay illustrates this claim through an analogy based on Daniel Dennett’s warning to prospective philosophy graduate students in “Higher-order Truths About Chmess.”.

Consider the following argument: All men are mortal; Socrates is a man; therefore, Socrates is mortal. Intuitively, what makes this a valid argument has nothing to do with Socrates, men, or mortality. Rather, each sentence in the argument exhibits a certain logical form, which, together with the forms of the other two, constitute a pattern that, of itself, guarantees the truth of the conclusion given the truth of the premises. More generally, then, the logical form of a sentence of natural language is what determines both its logical properties and its logical relations to other sentences. The logical form of a sentence of natural language is typically represented in a theory of logical form by a well-formed formula in a ‘logically pure’ language whose only meaningful symbols are expressions with fixed, distinctly logical meanings (e.g., quantifiers). Thus, the logical forms of the sentences in the above argument would be represented in a theory based on pure predicate logic by the formulas ‘∀x(Fx ⊃ Gx)’, ‘Fy’, and ‘Gy’, respectively, where ‘F’, ‘G’, and ‘y’ are all free variables. The argument’s intuitive validity is then explained in virtue of the fact that the logical forms of the premises formally entail the logical form of the conclusion. The primary goal of a theory of logical form is to explain as broad a range of such intuitive logical phenomena as possible in terms of the logical forms that it assigns to sentences of natural language.

A formal, computational, semantically clean representation of natural language is presented. This representation captures the fact that logical inferences in natural language crucially depend on the semantic relation of entailment between sentential constituents such as determiner, noun, adjective, adverb, preposition, and verb phrases.The representation parallels natural language in that it accounts for human intuition about entailment of sentences, it preserves its structure, it reflects the semantics of different syntactic categories, it simulates conjunction, disjunction, and negation in natural language by computable operations with provable mathematical properties, and it allows one to represent coordination on different syntactic levels.

Chomsky has constructed an empirical theory about syntactic universals of natural language by defining a class of possible languages which includes all natural languages (inter alia) as members, and claiming that all natural languages fall within a specified proper subset of that class. I extend Chomsky's work to produce an empirical theory about natural-language semantic universals by showing that the semantc description of a language will incorporate a logical calculus, by defining a relatively wide class of possible calculi, and by specifying a proper subset of that class which, I hypothesize, includes the calculi needed for the semantic description of any natural language. I argue that the special status, with respect to natural languages, of this particular type of logical calculus is an empirical finding which does not follow from any independently-known principles, and I conclude that the question why the laws of human thought have the structure they do is a biological rather than a logical question.

Is logic, feasibly, a product of natural selection? In this paper we treat this question as dependent upon the prior question of where logic is founded. After excluding other possibilities, we conclude that logic resides in our language, in the shape of inferential rules governing the logical vocabulary of the language. This means that knowledge of (the laws of) logic is inseparable from the possession of the logical constants they govern. In this sense, logic may be seen as a product of natural selection: the emergence of logic requires the development of creatures who can wield structured languages of a specific complexity, and who are capable of putting the languages to use within specific discursive practices.

Logic has its roots in the study of valid argument, but while traditional logicians worked with natural language directly, modern approaches first translate natural arguments into an artificial language. The reason for this step is that some artificial languages now have very well developed inferential systems. There is no doubt that this is a great advantage in general, but for the study of natural reasoning it is a drawback that the original linguistic forms get lost in translation. An alternative approach would be to develop a general theory of the natural logic behind human reasoning and human information processing by studying formal logics that operate directly on linguistic representations. That this is possible we will try to make plausible in this paper. It will turn out that one level of representation, that of Logical Form, can meaningfully be identified with the language of an existing and well-understood logic, a restricted form of the theory of types. It is not difficult to devise inference systems for this language, and it is thus possible to study reasoning systems that are based directly on language.

Though, at first sight, logical formalization of natural language sentences and arguments might look like an unproblematic enterprise, the criteria of its success are far from clear and, surprisingly, there have only been a few attempts at making them explicit. This paper provides a picture of the enterprise of logical formalization that does not conceive of it as a kind of translation from one language (a natural one) into another language (a logical one), but rather as a construction of a 'map' of (a piece of) the 'inferential landscape' of the natural language. The criteria that appear to govern the enterprise are labeled as those of reliability, ambitiousness, transparency and parsimony. These criteria, it is argued, do not provide for an excavation of a ready-made logical structure, but rather help us achieve a "reflective equilibrium" between the normative authority of logic and the answerability of logic to a natural language.

The relationships between logic and natural language are multiverse. On the one hand, logic is a theory of argumentation, proving and giving reasons, and such activities are primarily carried out in natural language. This means that logic is, in a certain loose sense, about natural language. On the other hand, logic has found it useful to develop its own linguistic means which sometimes in a sense compete with those of natural language. This has led to the situation where the systems of logic can be taken as interesting "models" of various aspects of natural language. Â Â Â Â Â Â Â The alliance of logic and linguistics has flowered especially from the beginning of the seventies, when scholars like Montague, Lewis, Cresswell, Partee and others showed how semantics of natural language can be explicated with the help certain suitable logical calculi and the corresponding model theory. (Montague went so far as to claim that in view of this, there is no principal difference between natural and formal languages - but this is, as far as I can see, rather misguiding.) Since that time, the interdisciplinary movement of formal semantics (associating not only linguists and logicians, but also philosophers, computer scientists, cognitive psychologists and others) has yielded a rich repertoire of formal theories of natural language, some of them (like Hintikka's game-theoretical semantics or the dynamic logic of Groenendijk and Stokhof) being based directly on logic, others (like the situation semantics of Barwise and Perry or DRT of Kamp) exploiting different formal strategies. Â Â Â Â Â Â Â Moreover, although the enterprise of formal semantics (i.e. of modeling natural language semantics by means of certain formal structures) seems to be the principal point of contact between linguistics and logic, there are also other cooperative enterprises. One of the most fruitful ones seems to be the logical analysis of syntax, which has resulted from elaboration of what was originally called categorial grammar. (However, even this enterprise can be seen as importantly stimulated by Montague.) Â Â Â Â Â Â Â All in all, the region in which logic and theoretical linguistics overlap has grown both in size and fertility..