We look at coordinated objects with different adverbs applying to each, and at adverbs before the verb.
Coordinated direct objects
(195) Nero gave Olivia gladly and Quintia sadly to Poppaea
In (195) the adverbs are comprehended as reflecting the attitudes of AGENT to the giving, not the attitudes of THEMEs. The attitude is however different for each THEME. It is not possible for that to be conveyed by GIVE / MANNER / GLAD and GIVE / MANNER / SAD because these are contradictory and fail to identify theme with attitude.
The proposal is that the relevant junction in this situation is not (verb)__(adverb) but the closer (noun)__(adverb). In (195) for example, the junction Olivia__gladly creates (null) / OLIVIA / GLAD / 6 and this forces consolidation of OLIVIA as THEME.
Several principles are involved here:
- activation is brought by the dependent
- the junction gave__Olivia results in a 4/2 split of activation
- incomplete propositions are unstable and a pair of them always tries to form at least one complete proposition (see LS33).
Use of a concept like OLIVIA as a relation has not been seen before. GLAD is how an AGENT may relate to its verb and that purpose is achieved generally no matter how the concept is activated. OLIVIA is how specifically GLAD and the verb are related in this instance.
QUINTIA is treated similarly. Then, as shown in (195), QUINTIA is coordinated with OLIVIA, and POPPAEA is GOAL for both THEMEs.
Coordinated indirect objects
Differences of AGENT attitude can also apply to GOAL.
(196) Nero gave Olivia to Poppaea gladly and (to) Renata sadly
In (196) the relevant dependencies are Poppaea__gladly and Renata__sadly. The resulting propositions – (null) / POPPAEA / GLAD / 6 and (null) / RENATA / SAD / 6 – do not force any consolidation because theme and goal propositions are already complete. LS35 explains how the correct propositions are delivered in a sequence that doesn’t mislead cognition.
Sequence matters here. LS8’s idea of an unsequenced bundle of propositions can be preserved if we assume the bundle applies to a single clause, the Renata phrase forming a second clause with ellipsis for which the AGENT and THEME propositions are carried forward.
The analysis is less obvious where the adverbs apply to GOAL but each one precedes its PP.
(197) Nero gave Olivia gladly to Poppaea and sadly to Renata
Now there is a complete sentence at Poppaea. At that point, Olivia__gladly has created (null) / OLIVIA / GLAD which, becoming GIVE / OLIVIA / GLAD, correctly represents AGENT attitude but unnecessarily attributes it to OLIVIA. Subsequently sadly links way back to gave. This is confirmed by an ungrammatical variant of (197).
(198) * Nero gave Olivia gladly to Poppaea and sadly Renata
The junction gave__sadly changes the C for gave and prevents bare-noun Renata following – as for sentences (5d), (7d) etc in LS51. However in (197), the final PP can still be accommodated giving a second goal proposition. From this second clause, only the more general GIVE / MANNER / SAD is created – from gave__sadly – and once again cognition must get the intended meaning by virtue of the sequence in which the propositions are delivered.
LS51 assumed that the adverb is always in a (verb)__(adverb) junction. Coordinated objects show that generalisation needs changing to a (preceding word)__(adverb) junction. In sentences such as (12d) and (20d) the adverb immediately follows the verb and the junction is still gave/given__gladly which creates GIVE / MANNER / GLAD / 6 as shown before.
Before the verb
An adverb before the verb presents a different set of issues.
The following active sentences have …Olivia to Poppaea after the verb. Equivalents with …Poppaea Olivia are assumed to behave in the same way.
(199) Nero gladly gave Olivia to Poppaea
(200) Gladly Nero gave Olivia to Poppaea
These are both good although (200) sounds quaint. Passives are not so clear-cut.
(201) Olivia was gladly given to Poppaea by Nero
(202) Poppaea was gladly given Olivia by Nero
(203) ?? Gladly Olivia was given to Poppaea by Nero
(204) ?? Gladly Poppaea was given Olivia by Nero
(203) and (204) are quite bad. These and similar examples would be perfectly grammatical with sentence-initial gladly replaced by sadly. The latter can have whole-sentence scope instead of qualifying a particular noun. In that case sadly would be equivalent to It was sad that but Nero need not have been sad at all. This effect is possible with any adverb that can express SPEAKER’s attitude rather than AGENT’s.
In (203) and (204), gladly really applies to remote agent rather than adjacent THEME or GOAL. Omitting by Nero perhaps worsens these sentences. This is borne out by using a temporal (and not strongly AGENT-related) adverb rather than a manner adverb in this construction.
(205) Yesterday Olivia was given to Poppaea (by Nero)
(206) Yesterday Poppaea was given Olivia (by Nero)
(205) and (206) are perfectly good. Sentence-initial gladly with the same auxiliary but in an active construction is still a bit odd.
(207) ? Gladly Nero was giving Olivia to Poppaea
Placing the adverb between subject noun and auxiliary doesn’t have much effect.
(208) ? Nero gladly was giving Olivia to Poppaea
(209) Nero yesterday was giving Olivia to Poppaea
It would seem that gladly__gave and yesterday__was exist in language knowledge, but gladly__was doesn’t. Perhaps that is because copular be forms don’t work with gladly.
(210) * Gladly he was old
The intended meaning in (210) must be expressed differently.
(211) He was glad to be old
But copular be does work with yesterday.
(212) Yesterday he was glad
Even if language knowledge doesn’t have gladly__was, it might have gladly__giving/given. But that should make (203), (204), (207) and (208) unexceptionable, which is clearly not the case.
The behaviour of gladly before the verb in active sentences with a fronted object provides clues to what is behind these difficulties.
The discussion is simplified by basing the examples on Olivia Nero gave to Poppaea and assuming that essentially the same behaviour applies to Poppaea Nero gave Olivia.
(213) Olivia Nero gladly gave to Poppaea
(214) * Olivia gladly Nero gave to Poppaea
(215) * Gladly Olivia Nero gave to Poppaea
In (213), if gladly__gave were processed first, it might be that GIVE / MANNER / GLAD / 6 is delivered immediately and doesn’t interfere with the not-yet-created propositions for the valents of GIVE. However, if yesterday is substituted for gladly, all three sentences are acceptable: yesterday__gave works with yesterday anywhere before the verb, with Nero and Olivia handled as in sentence (13). Perhaps the difference is that gladly is a manner adverb and has a semantic association with AGENT, while yesterday is a temporal and applies more generally to the verb. How then is gladly handled?
If there is affinity between gladly and an adjacent noun, (213) gives Nero__gladly, (214) gives Olivia__gladly and (215) gives gladly__Olivia. The first two of these dependencies seem to be of a type already seen where the adverb occurs after the verb. In (195) for example, Olivia__gladly creates (null) / OLIVIA / GLAD / 6.
That proposition alone does not account for the ungrammaticality because it could coexist with GIVE / THEME / OLIVIA – as happens in (195). The proposal is that, because a pre-verb manner adverb is associated with AGENT and because the null must eventually be replaced by the concept for the verb, Olivia__gladly creates two incomplete propositions. The expectation is that the other proposition will be activated when a Olivia__(verb) junction is encountered.
This is incompatible with the frame created by Nero__gave and the ungammaticality occurs.
It might be objected that (noun)__(adverb) is creating different structures depending on whether it occurs before or after the verb. This can be countered by again distinguishing the C for a post-verb noun because it has already participated in (verb)__(noun).
Loose-ends in this account are tied up if, before the verb, Olivia__gladly and gladly__Olivia create exactly the same proposition and Nero__gladly and gladly__Nero both create (null) / NERO / GLAD / 6. Then the analysis for (215) is exactly the same as that for (214). (213) – and also (199) and (200) – are grammatical because early creation of (null) / AGENT / NERO is perfectly compatible with the propositions for Nero__gave. (203) and (204) are ungrammatical because early creation of (null) / AGENT / OLIVIA (or POPPAEA) is not compatible with the propositions for the valents of passive GIVE.
(208) and (209) create a similar pair of propositions from their first junction.
The doubtful grammaticality of (208) evidently relates to the interaction of this structure with the propositions for Nero__was, which is as for Nero__is in (5). It should be possible to merge the proposition from the latter with (null) / AGENT / NERO / 0. But forming something like BE / NERO / GLAD / 6 seems to be impossible despite gladly__was – as in (210) – being avoided (because gladly can’t be a dependent more than once). Presumably (null) / NERO / GLAD / 6 cannot be held over until giving is encountered.
Apologies. The treatment of auxiliaries has not been elaborated in LanguidSlog but I hope the reasoning here, although difficult, will prove robust.
This piece has become far too long because it’s been difficult to find a break point. Next time we’ll finish with adverbs by looking at coordinated subjects and at probably.