The primary aim of the Tractatus is to find an explanation for the possibility of a proposition's having sense. Behind this stands the presupposition of the Tractatus that only the proposition can correspond to something in reality, or, if it is true, does correspond to it. "Only the proposition has sense …" (TLP 3.3), and it presents a state of affairs (TLP 3.122, 3.1232). The truth of a proposition is contingent; the possibility of its truth (of its content), that is, its sense, is guaranteed a priori by its form (TLP 3.13). In contrast to the proposition, names have meaning (TLP 3.203). Meaning cannot be true or false (TLP 3.221) but must be presupposed so that a proposition can be true or false (i.e. significant): names mean possibilities of sense.

How, then, is the possibility of the significant proposition, e.g. the understandability of sense, to be explained? The answer of the Tractatus is that the sense of the proposition is understood when the meanings of its constituents are understood (TLP 4.024). So the meanings of the constituents of the proposition must be presupposed as given.

Against this stands the fact that the meaning of a propositional constituent can itself only be understood as its application in propositions in which it occurs (or can occur):

"Only the proposition has sense; only in the context of a proposition has a name meaning." (TLP 3.3)

This is expressed quite decidedly in connection with learning and teaching ("explaining") meaning by means of elucidations:

"The meanings of primitive signs can be explained by elucidations. Elucidations are propositions which contain the primitive signs. They can, therefore, only be understood when the meanings of these signs are already known." (TLP 3.263)

"The meanings of the simple signs (the words) must be explained to us, if we are to understand them. By means of propositions we explain ourselves." (TLP 4.026)

But how could we make ourselves understood by means of propositions if we only come to know the meanings of their constituents through elucidations, by means of which we, in turn, could only make ourselves understood if we understand their constituents? The programme of a theory of sense, to explain the possibility or understandability of a significant proposition, falls into the following dilemma: the understandability of (the sense of) a proposition and the understandability of (the meaning of) its constituents are interlocked in a way that makes it impossible to take the one as the presupposition of the understandability of the other. Since both moments appear on the one hand as instances to be grounded and on the other hand as grounding instances, and are thus evidently arranged in a self-referential figure, no absolute starting point of grounding is in sight. How, then, is one to get behind the self-reference between the understandability of sense and the understandability of meaning?

One can neither get behind it nor fail to get behind it; rather, to continue the metaphor, one simply has to go along with it. In the use of language, the competent speaker travels the self-referential path from "sense" to "meaning" and from "meaning" to "sense". Depending on the direction, I call this path formalization or de-formalization. To the respective stages of formalization there correspond "expressions" or "formal concepts"; they express a possibility of sense. An expression is the "common characteristic mark of a class of propositions" (TLP 3.311), presented by a propositional variable "whose values are the propositions which contain the expression" (TLP 3.313). The (de-formalized) proposition, on the other hand, is the composition, the "function of the expressions contained in it" (TLP 3.318). In the Tractatus, formalization is presented in its genuinely logical-analytic context as follows:

"If we change a constituent part of a proposition into a variable, there is a class of propositions which are all the values of the resulting variable proposition. This class in general still depends on what, by arbitrary agreement, we mean by parts of that proposition. But if we change all those signs, whose meaning was arbitrarily determined, into variables, there is always still such a class. But this is now no longer dependent on any agreement; it depends only on the nature of the proposition. It corresponds to a logical form, to a logical prototype." (TLP 3.315)

An example: the proposition (1) "The house is red", since it is not applied here as a linguistic act, is already the result of a pragmatic step of formalization, i.e. a propositional variable whose values are all its situationally significant de-formalizations. A further semantic or syntactic step of formalization yields the propositional variables (2) "The house is colour-determinable" and (3) "x is red". (2) means the possibility of the colour-determinable object house; (3) is a predicator meaning the possibility of the determination red of colour-determinable objects. As the result of a further, again semantic or syntactic, step of formalization, there follow the propositional variables (4) "The house is determinable" and (5) "x is colour-determinable". (4) is the nominator house, which means the possibility of the in-some-way-determinable object house; (5) is a logical form containing the possibility of the colour-determinability of objects. In the last step of formalization one obtains the propositional variable (6) "x is determinable", the formal concept of the nominator or predicator as such, which means the possibility of the in-some-way-determinability of objects.

This now permits an explanation of the theory of the composition of the proposition underlying the Tractatus. In principle, two levels of propositional composition can be distinguished: that of the de-formalized (actualized) proposition, which expresses a sense, and that of the formalized proposition (propositional variable), which expresses a meaning (a possibility of sense). A propositional constituent is accordingly a propositional variable expressing a possible composition of propositions (a possibility of sense). And the proposition is not a material connection of expressions which per se mean factual entities, but the composition of possible propositional contexts (propositional variables), namely as their de-formalization, and vice versa, in the course of the propositional act.

Further, this conception contradicts those interpreters who reproach the Tractatus with leaving out "reflection". The formalization of the sense of propositions and actions in the direction of the possibilities of sense underlying them is in fact an equivalent of "reflection". This holds both for the implicit reflection that must be presupposed for linguistic competence and for explicative (e.g. transcendental-philosophical) reflection, which, by formalizing and systematizing a language-action game, shows its possibilities of sense to be its conditions of possibility and explicates them as a language-action system. (De-)formalization is thus the fundamental conditio sine qua non of the significant propositional act and of language-action competence, and secondarily also the conditio sine qua non of philosophy (theory of sense) as the systematic explication of language-action competence.

The concept of (de-)formalization also makes a (radical-)constructivist interpretation of the Tractatus plausible. Reality can be defined as reality describable in a system; essential to describable reality, i.e. to reality as a "domain of description" (Maturana 1982, p. 76), is the "isomorphism of the logic of the description and the logic of the describing system" (ibid.). In the terminology of the Tractatus: "What the picture must have in common with reality in order to be able to represent it after its manner – rightly or falsely – is its form of representation." (TLP 2.17; cf. TLP 4.12 and elsewhere). Against the view that the Tractatus starts from a given reality in itself (and so holds a position diametrically opposed to Radical Constructivism), there speaks a series of propositions in which the primary sense-constituting role is ascribed to "application" (synonymous with "construction") (cf. TLP 3.263, 3.5, 3.326 ff., 5.557).³