For frege, To define a symbol is to show how to do without it. Frege originated the distinction between metalanguage and object language. But he formulates his definitions within the begriffsschrift itself, Not seeing that, According to his own account of them, They go better in the metalanguage.

Any semantic theory is bound to presume some structure in the messages it analyses, and the success of the theory depends on getting this structure right. But discovering this structure is the business of grammar. Therefore grammar is a necessary preliminary to semantics. Semantic theories of conditionals vividly illustrate this. All presume a provably untenable ternary structure: antecedent, operator, consequent. And all can be shown committed as a result to a thoroughly unbelievable set of connections between sentences and their informational (...) burdens. Actually, a conditional has five immediate factors, none of them an antecedent or a consequent. (shrink)

It is urged (1) that geach is correct in his claim ("assertion", "philosophical review", 74, (1965), Page 449) that what he calls 'the frege point' is logically independent of frege's doctrine that sentences are names of objects, And (2) that frege's 'propositions of begriffsschrift' are neither truths nor falsehoods.

Received doctrine has an 'antecedent' message encoded within a conditional clause, such as the string comprising the first five words of the sentence 'If the bough had broken the cradle would have fallen'. Criticisms of mine of this tenet were recently challenged by Stephen Barker. In the course of responding to his examination, I venture a snappy demonstration that the 'conditionals' such sentences encode can have neither 'antecedents' nor 'consequents'. Also, less happily, I urge a binary outermost structure for these (...) conditionals. (shrink)

This brief note shows that the following three tenets, All to be found in frege's "on sense and reference", Form an inconsistent triad: (1) two proper names express the same sense iff their identity-Sentence "contains no actual knowledge"; (2) sentences are proper names; (3) if in a sentence we replace one proper name by another having a different sense, "we see that in such a case the thought changes".

It is argued that it is a misinterpretation of frege to construe his talk of the bedeutungen of sentences and of predicates in an ontologically innocent way.

Book reviewed in this article:F.H. Bradley, Collected Works Volumes 1–5.

The commonplaces, all grammatically confused, are that ?conditionals? are ternary in structure, have ?antecedents? and conform to the traditional taxonomy. It is maintained en route that ?The bough will not break? is consistent with ?If the bough breaks ??, that there is no logical difference between ?future indicatives? and ?subjunctives?, and that there is a difference between the logic of propositions (e.g. ?The bough broke?) and that of judgments (?The bough will/might/could/should/must/needn't break?).