On floating, raising, and being
The distribution of floating quantifiers (FQs) has formed an important diagnostic for intermediate subject positions. Sportiche (1988) proposes that FQs are generated together with the subject and can then be stranded when the subject is moved. However, it is not the case that FQs can occupy any position that we expect the subject to have passed through. For example, FQs are not allowed immediately after the auxiliary being in a progressive passive sentence (see Harwood, 2011; Bošković, 2013; and references therein):
(1) We could (✓all) have (✓all) been (✓all) being (*all) punished (*all) for our crimes (*all). (based on Harwood, 2011)
Previous work has proposed to account for this pattern by requiring that FQs adjoin to the subject only after a certain point in the derivation. For example, for Bošković (2004), FQs cannot be adjoined to an argument in its theta-position; for Harwood (2011), an FQ cannot be adjoined until the end of the construction of a phase. Once the FQ has adjoined to the subject, it can be stranded in any higher position that the subject moves through. These approaches therefore predict a bifurcation, where FQs can appear above a certain point, but not below it.
Not discussed in this context, however, are cases of passives of ECM verbs, which function as subject-raising predicates. In these cases, FQs are rejected after being in the higher clause, but are allowed in the lower nonfinite embedding:
(2) The students are (✓all) being (*all) {expected/predicted} to (✓all) finish their dissertations in time.
One complication affecting examples such as (2) is that some speakers judge progressives of ECM passives and raising predicates more generally as marked. However, naturally occurring examples show that such constructions are possible and that they are indeed raising constructions, as diagnosed by idiom chunk interpretation:
(3) “And of course the shit could hit the fan at any time - but the shit is being predicted to hit the fan constantly and has been for years.” (naturally occurring example from http://www.davidicke.com/forum/showthread.php?p=1060803118)
Example (2) is an important data point for analyses of FQ. The distribution of FQ in (2) is unexpected under a view where, once an FQ is merged, it can be stranded in any higher intermediate position. The inability of FQs to follow being in (1–2) therefore must be due to an independent ban on FQ in this post-being position. Assuming the basic stranding approach to FQ (Sportiche, 1988), we learn that this position is special in that either (a) the subject does not land in an intermediate position immediately following being or (b) FQs are not allowed to be stranded there.
References:
Bošković, Željko. 2004. Be careful where you float your quantifiers. Natural Language and Linguistic Theory 22. 681–742.
Bošković, Željko. 2013. Now I’m a phase, now I’m not a phase: On the variability of phases with extraction and ellipsis. Linguistic Inquiry 45:1.
Harwood, Will. 2011. There are several positions available: English intermediate subject positions. Proceedings of ConSOLE XIX.
Sportiche, Dominique. 1988. A theory of floating quantifiers and its corollaries for constituent structure. Linguistic Inquiry 19. 425–449.