2.4 Ditransitives and double objects
2.4.2 Difference in meaning between DAC and DOC
2.4.2.2 The status of the external argument and the direct object
So far the difference between DAC and DOC was attributed only to the status of the indirect object. Now, there is an important difference between the two constructions which has to do with the status of the external argument in each of them. Consider the following pair, noted originally by Oherle (1976):
(40) a. Nixon gave a book to Mailer, b. Nixon gave Mailer a book.
As noted by Oherle, (40b) is ambiguous between two readings. On the first reading Nixon literally gave a book to Mailer (with or without a change o f possession). On this reading Nixon is a volitional agent, and the book he gives is a physical object. The second reading is paraphrased by Oherle as "Mailer wrote a book which he wouldn't have been able to write if it hadn't been for Nixon". On this reading Nixon is not an agent which performs an act of giving, and the book is not a physical object that he hands to Mailer.
The important observation is that the second reading arises only with the double object construction, never with the dative construction:
(41) a. The war years gave Mailer his first big success. b. *The war years gave his first big success to Mailer. c. Interviewing Nixon gave Mailer a book.
d. *Interviewing Nixon gave a book to Mailer. (Oherle 1976, Pesetsky 1995: 193)
When the subject is inanimate (i.e., it cannot be interpreted as a volitional agent), only the DOC is allowed:
(42) a. The flowers gave the room a festive atmosphere. b. *The flowers gave a festive atmosphere to the room.
(43) a. Katya taught me Russian. b. Katya taught Russian to me.
c. Lipson's textbook taught me Russian.
d. *Lipson's textbook taught Russian to me. (Oherle 1976)
Ditransitives are often used as psychological verbs in English (see chapter five for discussion). As expected, only the DOC, in which the subject need not be an agent, is a well-formed psych verb:
(44) a. The trial gave her a lot o f grief. b. *The trial gave a lot o f grief to her. c. Paul / The noise gave Mary a headache. d. *Paul / The noise gave a headache to Mary.
DOC include many forms which can be paraphrased as simple transitive verbs:
give a kiss (i.e. kiss), give a kick (i.e. kick) etc. In all such cases, the thing given is not a real object, but a figurative one:
(45) a. Paul gave the baby a bath. b. *Paul gave a bath to the baby.
(45b) cannot mean Paul bathed the baby. The only way to interpret it is, that Paul gave a bath (an object) to the baby. Some linguists have argued that give acquires a causative
sense on the DOC. Its meaning is something like "cause x to have y " (Pinker 1989, Goldberg 1995, Pesetsky 1995):
(46) a. John gave Paul a kiss / a kick / an insult. b. *John gave a kiss / a kick / an insult to Paul.
c. Cardamon pods gave the pudding a sharp, distinctive taste. d. * Cardamon pods gave a sharp, distinctive taste to the pudding.
I will assume here that the indirect object may be merged at two possible positions: one is above the direct object, as in DOC. At this position it is interpreted as "affected" by the action. The second position is below it, as in DAC. This position is also the one occupied by locatives (e.g. put a book on the shelf). The indirect object is interpreted there as a goal, towards which the action is directed (but not necessarily affecting him directly). The structures I suggest for both constructions are as follows:
(47) a. DAC vP Agent vP DOC X ? NP Theme VP V V PP Location Agent JCP VP NP Goal V V V NP Theme
X is the head of the upper projection in the case of DOC. It may be v, if the verb is agentive, or another head, if it is not (see chapter five). With DAC, the VP necessarily includes v which projects the agent. I termed the two positions o f the indirect object as "Goal" (the higher one) and "Location", for convenience. The higher position o f the indirect object is the position which is associated with some "affectedness" o f this object, as noted above. The direct object may be interpreted as "being acted towards" or "being affected by the action", depending upon where it is generated.
The hypothesis about two positions for indirect objects is supported by the fact that even in Hebrew, which does not have double objects but allows the indirect object either to precede or to succeed the direct object, similar interpretational effects arise:
(48) a. Hi sipra et ha sipur le Tom (aval hu lo hikshiv) she told OM the story to Tom (but he didn't listen) b.
hikshiv)
Hi sipra le Tom et ha sipur (*/??aval hu lo
she told to Tom OM the story (*/?? but he didn't listen)
When the indirect object occupies the higher position (48b) it is interpreted as affected, while at the lower position it may be interpreted as not fully affected.
To conclude the discussion of DAC and DOC I will sum up the main points I made:
1. There exists a real difference in interpretation between DAC and DOC. This makes a transformational account less plausible.
2. There seem to be two positions for indirect objects: one lower than the direct object, shared with locatives; the other is above it.
2. With DOC the indirect object is necessarily interpreted as affected and the event is interpreted as fully "transmitted". With DAC no such requirement exists.
3. With DOC the subject may be interpreted as non-agentive and the direct object may be interpreted not as a physical object, and the act o f giving may be interpreted as "contact" rather than volitional transference. With DAC, the subject is always agentive, the object is always physical and the act o f giving is volitional transference.