Language-Game, Language System, Language Ideal
On the Possibility of a Self-Referential Theory of Sense following Ludwig Wittgenstein
First published as "Sprachspiel, Sprachsystem, Sprachideal. Über die Möglichkeit einer selbstreferentiellen Sinntheorie im Anschluß an Ludwig Wittgenstein", in: Theo Hug (ed.): Die soziale Wirklichkeit der Theorie. Beiträge zur Theorievermittlung und -aneignung in der Pädagogik, Munich (Profil) 1990, pp. 237–249.
The philosophy of Ludwig Wittgenstein lends itself particularly well as a forum for discussing current philosophical questions. This applies not only to the Wittgenstein of the Philosophical Investigations, by now cited extensively in linguistics and the social sciences, but also to the Wittgenstein of the Tractatus Logico-Philosophicus. There, as is to be shown, a theory of language as a self-referential system is laid down which in part bears features of so-called Radical Constructivism and which (with a view to the Philosophical Investigations) announces the postmodern awareness of the "heterogeneity of language-games". The Tractatus shows that the "project of modernity", to legitimate language and reality by recourse to a universal language system, fails, and must fail from within.1 As a language system presupposed to be universal, it is immune to external objections; it would have the omnipotence to define what is meaningful and valuable. It becomes entangled in internal contradictions, however, when it does so, i.e. when it attempts to explicate and legitimate its own system as universal. The alternative the Tractatus leaves behind is either to remain silent or to give up the assumption of a universal language system.2 In what follows, the line of argument is traced from the Tractatus to the Philosophical Investigations, where this assumption is abandoned and a new kind of practice of legitimation is sketched.
The guiding thread is the question of the possibility and programme of a self-referential theory of sense. By "theory of sense" I mean the reconstruction of the possibilities of sense that must be presupposed for the sense of propositions and actions. The aim of a theory of sense is the systematization of a language-action game, that is, the reconstruction of the underlying language-action system. Such a reconstruction is possible insofar as for every language-action game alternative language-action games are conceivable: the totality of these alternatives I call the language-action ideal. A system is the result of reflexively going beyond a game within the horizon of the ideal.
A theory of sense is "self-referential" insofar as meaning and understanding are possible on the basis of a self-referential connection between possibilities of sense and the significant propositional act. In what follows, this circular connection is specified as formalization and de-formalization: a possibility of sense is the formalization of a significant propositional act; a significant propositional act is the de-formalization of a possibility of sense. The self-regulation of formalization and de-formalization is essential to the possibility of the propositional act's being significant.
With regard to the possibility of such a theory of sense, the Tractatus holds an ambivalent position. Although a theory of sense is practised in the sense indicated, the impossibility of an explanatory theory of sense is (precisely thereby) shown: there is only the explicative "activity" as an "elucidation", appealing to examples, of possibilities of sense that "show" themselves (cf. TLP 4.112 and elsewhere). This position of the Tractatus has a weakness and a strength. The weakness is that it involves the fundamentalist assumption that the possibility of language as a whole is to be limited to the language system of the depicting-asserting language-game. Under this assumption (as the Tractatus shows), the possibility of the theory of sense itself cannot be contained in the possibility of the language system it thematizes without infinite regresses and paradoxes arising. The strength of this position is that, in the second instance, it shows the untenability of the assumption of a universal language system, since it is precisely and only under this assumption that a theory of sense is impossible.