Cross-linguistically, very few complete sentences, as opposed to a myriad of phrases, lexicalize to become words.
I here offer an account for this skewed distribution, along the lines of Construction Grammar, by analyzing a set of mono-clausal
sentences in Hebrew which have indeed become – or are on the verge of becoming – words. I adopt the distinction between
categorical and thetic propositions, and show that only the latter can evolve into words. A thetic – unlike a categorical –
proposition, much like a verb-phrase, enables a tight semantic bonding between its components to form an ‘interpretatively
cohesive’ unit, which may lead to semantic change. An evaluative thetic – unlike a categorical – proposition is comment-like,
hence ‘semantically-incomplete’, and in need of a topic from prior discourse to predicate on, which may lead to a change in
grammatical status. All verb-phrases meet these two criteria but only few sentences do, hence, I argue, the skewed distribution of
sources from which new words evolve.