V83.0076: More on the Language of Probability

First appearances suggest that probability operators (is as likely as, probably, is more likely than, etc) should be treated together with what we might call pure modal talk (might, may, ought, must, possibly, necessarily, etc) because: • They are non-truth-functional operators on propositions. • They interact logically with epistemic modals: R3. Must to probably. It must be that Probably R4. Probably to might. Probably It might be that/It is possible that You may, however, have noticed that: • Comparative forms are not as easy/natural with straight modal talk: ? is as possible as, ? is as necessary as. Hopeless with modal auxiliaries. • The various comparative probability operators (is as likely as, is (much) more/less likely than, etc) obviously have a bunch of a further decomposable structure. The range of morphology involved in this further structure occurs in comparative discourse generally (this is what Kennedy [2006] calls degree morphology): comparative morphemes: more, less, as intensifiers: very, quite, rather su ciency morphemes: too, enough, so and obviously, all this morphology is playing the same semantic role. That last observation tells us very clearly that our semantics for probability operators is really going to have to play nice with the true semantics for degree morphology.