Atomic Facts, Propositions, Proposals

In the Tractatus, Wittgenstein defines an 'atomic fact' as a "combination" of simple objects (#2.01-02), but without specifying what those objects are, e. g. minds, sense-data, etc.  Similarly, his model of Language is based on the Propositions that correspond to Atomic Facts, without a specification of which Propositions in, say, German or English, are the simple ones.  Now, for the most part, the relation of correspondence between a Proposition and a Fact in the Tractatus seems to be that of Description, i. e. he regularly characterizes the former as a "picture" of the latter.  However, a subtle, but important, alternative is suggested by #4.031--"In the proposition, a state of affairs is, as it were, put together for the sake of an experiment"--which formulates a Constructivist concept of Language, i. e. the passage implies that even an Atomic Fact is "put together" by the Proposition which corresponds to it.  Furthermore, "for the sake of experiment" indicates that the sequence of words that 'puts together' the facts of the world is, more precisely, a Proposal, not a Proposition.  Accordingly, all the components of the Language of the Tractatus, starting with the pictures of Atomic Facts, are Proposals.  Whether or not Wittgenstein recognizes that he is advocating Constructivism in the Tractatus is unclear.