Quirk’s thorough treatment of of, though falling mainly outside of the chapter on prepositions, is invaluable, and in this case provides what I think is a more reliable analysis of of’s functionality than the ODE sense inventory does. The main treatment is in ¶17.40ff; reading these sections thoroughly rewards the study of the preposition.
A slight peculiarity of the ODE sense division is that, while most senses are defined semantically, senses 6 and 7 (including their subsenses, making 6 senses in all and corresponding to DIMAP 11 through 16) are defined grammatically. This makes some overlap inevitable, in that certain of phrases correspond semantically to a particular ODE sense, but also meet the qualifications of one of the grammatical patterns described in senses 6 and 7. In these cases I have assigned sentences to the sense that I thought was more specific while still being accurate — viewing the senses defined grammatically as least specific. In other words, I have used Dimap senses 11 through 16 only for examples that (1) fit the pattern, and (2) didn’t fall more specifically under a semantically defined sense. Many more sentences could have been assigned to senses 11 through 16 than actually are, since they are rather large bins.
Notes from the spreadsheet:
New subsense defined: ODE has no obvious place to put a very common of usage, where of it follows the name of a container, and is complemented by the container’s contents, e.g., “bag of chips.” I have added this under sense 3(1b), since it is essentially partitive. Corresponds mainly to frame element contents.
New subsense defined: The phrases treated in ODE as be of corresponds mainly to Quirk’s ¶9.55, where of connotes having or possession, with an expressed or implied copula. These are actually fairly common and not labeling the FN instances of them would create a needless gap, so I have converted this phrase to a subsense — number 6(3)-1
NODE had one subsense of of (5a, 10 in DIMAP) that ODE absorbed into the major sense. I have retained this subsense since it can be usefully distinguished from the parent sense: the POA is always one of a limited number of words denoting broad categories: brand, kind, make, race, sort, strain, type, variety.
Other instances of for that may deserve idiomatic treatment
The ODE phrases sum it up; of them, only of all is likely to be encountered with any frequency.
There is a close relationship between senses 3(1b) (quantifier + partitive noun) and 17(8) (material or substance constituting something), and a sense division is conceivable in which these would fall under the same head. For example: hewing strictly to ODE’s wording, “pile of bricks” would fall under 17(8) and “pile of football players” under 3(1b) — since football players are presumably not classifiable as a material or substance. I have classified the FN element “individuals” all as 3(1b), though interestingly this element occurs in Frame Quantity (which suggests sense 3(1b)), as well as in Frame Aggregate (which to me equally suggests sense 17(8)).