Bucket of Grice

---

We are discussing with Kramer, etc., as to the ordering of things in "Gazdar's bucket". In his "Layers of Lift", etc.

As I noted in "Grice's bucket", the idea is Levinson's, I mean Gazdar's, I mean Grice's. I mean anyone! What it strikes the cliche-turned-adage is that it _is_ a bucket, yet one is supposed to throw things in an orderly fashion, and RETRIEVE things from the bucket in a similarly orderly fashion. Having fished quite a bit, and having thrown in buckets _ANTYHING_ in _Whatever_ order, I can't see what Levinson is meaning --. Plus, it's all defeasible, i.e the order holds, provided, as he writes: "they are all added to the bucket, provided the later additions are compatible with the previous ones."

Since he is an anthropologist and linguist (but not a philosopher) some of the nomenclature he uses will NOT amuse the philosopher, especially when uncredited, or taken slightly. He throws entailments first, but fails to credit Moore for the monstrosity in the first place!

"First entailments". So, first you are supposed to throw entailments. These are like meaning postulates, only different. They are if-clauses, but trickier. They are more like "logical implications" as Grice calls them. WoW:P&CI.

"Second, clausal Q GCIs." This are then implicata or implicatum of the non-scalar type. Followed by 'that'-clauses (E.g. "He knew that she was pregnant", you hear, so you do throw in the bucket, "She was pregnant", a clausal implicature. Gazdar invented them, but Urmson discovered them). The Q is supposed to merge the quantitative and the qualitative, in unGricean ways. And GCI is the monster of the generalised conversational implicature.

"Third, scalar ones". This are Horn's inventions, although Urmson does speak of 'scale' ("We can order 'believe' and 'know' in a scale"). The idea that in the ordered pair , provided there's an entailment relation to them which is truth-functional in terms of 'iff', the uttering of 'w' -- justifies you in throwing the negation of 's' into the bucket. "I think God exists" -- throw in the bucket: Utterer doesn't know.

"Fourth, M GCIs". This are manner or modal-implicatures, i.e. arising from the modal-maxims of the modal category of Mode. Why they come fourth should escape a modalist.

"Fifth, and last, I GCIs" "--with I and M implicatures the distinction between clausal and scalar ones is irrelevant)." and where "I" stands for "Informativeness", since Atlas and Levinson discovered (or thought they had discovered, that is) that the idea was lacking in Grice!

"They are all added to the bucket, provided the later additions are compatible with the previous ones."

--- Again, the incremental model is a good one. Personally, I'm more intrigued by more basic, philosophical questions, as developed in "Grice's Bucket", this blog. I.e. the steps in a working-out scheme of a rather general type, and how a clash of one maxim may lead to a further assumption, etc. The idea that it's all so orderly and -- in any case, thrown into an undistinguished bucket may be neither here nor there.

The practicalities of the Bucket escape me. I suppose the _term_ is misleading, "bucket". There must be a way of making the idea plausible in practical terms. Since my focus here has been the operators: &, v, and -->, I cannot see how bucket considerations will _kick_ it. Etc.