Academia.eduAcademia.edu

Figure 6 – uploaded by Rafael Marín

See full PDFdownloadDownload figure
The solution is to move the head Appl to v and displace both arguments, Goal and Theme, to spec, vP. In t configuration, the applicative and the two arguments are under the same constituent. Movement is triggered by fori requisites: goal and theme have to move to vP to get case assigned, checking features that Bruening (2010) hypothesi: are inherited from Voice. The unavailability of goals as externalized arguments with ditransitives amounts, thus, to why the configuration in (50), necessary for semantic interpretation, is impossible in an adjectival participle. Bruening suggests that when Adj

Figure 6 The solution is to move the head Appl to v and displace both arguments, Goal and Theme, to spec, vP. In t configuration, the applicative and the two arguments are under the same constituent. Movement is triggered by fori requisites: goal and theme have to move to vP to get case assigned, checking features that Bruening (2010) hypothesi: are inherited from Voice. The unavailability of goals as externalized arguments with ditransitives amounts, thus, to why the configuration in (50), necessary for semantic interpretation, is impossible in an adjectival participle. Bruening suggests that when Adj

Related Figures (6)

Contrasts like this meant, for Stowell, that these subjects are projected as external arguments (12a). A second exterr argument can be introduced — in a two-layered structure, with head-to-head movement of the lower head (12b), whi expresses the event where the behaviour is displayed and corresponding to Kratzer’s (1988/1995) spatiotempo variable. Sites St SOE ae aS The reader has probably noticed already that if several argument structures coexist for the same adjective — as in Stowell’s proposal, the claim is that adjectives can undergo argument alternations. Furthermore, if argument structure is essentially the same for verbs and adjectives, the prediction is that in adjectives we should be able to find alternations which are equivalent to those we find with verbs. Longobardi (1987) had already observed, for Italian, the alternation in (13), which Cinque (1990) considers parallel to the one in (14): there is an external argument role which can be assigned (as in 13a, 14a) or left suspended (13b, 14b).
Contrasts like this meant, for Stowell, that these subjects are projected as external arguments (12a). A second exterr argument can be introduced — in a two-layered structure, with head-to-head movement of the lower head (12b), whi expresses the event where the behaviour is displayed and corresponding to Kratzer’s (1988/1995) spatiotempo variable. Sites St SOE ae aS The reader has probably noticed already that if several argument structures coexist for the same adjective — as in Stowell’s proposal, the claim is that adjectives can undergo argument alternations. Furthermore, if argument structure is essentially the same for verbs and adjectives, the prediction is that in adjectives we should be able to find alternations which are equivalent to those we find with verbs. Longobardi (1987) had already observed, for Italian, the alternation in (13), which Cinque (1990) considers parallel to the one in (14): there is an external argument role which can be assigned (as in 13a, 14a) or left suspended (13b, 14b).
ee OE ES ee Several pieces of evidence support this conclusion. One of them is that, as Baker (2008: 48-49) points out, it explair why, cross-linguistically, adjectives cannot inflect for tense — once we eliminate, given independent tests, cases where inflected verb is cliticized to the adjective. Assuming that (i) tense inflection is combined with a lexical head through head-t head movement and that (ii) a functional head cannot be adjoined to tense, presence of Pred between A and T explains th: adjectives can never move to T. Movement to T skipping Pred violates the Head Movement Constraint (Travis, 1984); if stops in Pred, further movement of the complex to T is blocked, because Pred is a functional head (20a). Verbs do ni encounter this problem, because they do not need another head to introduce the subject (20b). Further empirical consequences are explored in Baker (2008). He starts from the observation that adjectives never truly agree in person with their subjects; other categories, such as v/V or T, can agree in person. He claims that the generalization can be captured by assuming that person agreement involves more rigid licensing conditions than number or gender, the Person Licensing Condition (PLC, Baker, 2008: 143).
ee OE ES ee Several pieces of evidence support this conclusion. One of them is that, as Baker (2008: 48-49) points out, it explair why, cross-linguistically, adjectives cannot inflect for tense — once we eliminate, given independent tests, cases where inflected verb is cliticized to the adjective. Assuming that (i) tense inflection is combined with a lexical head through head-t head movement and that (ii) a functional head cannot be adjoined to tense, presence of Pred between A and T explains th: adjectives can never move to T. Movement to T skipping Pred violates the Head Movement Constraint (Travis, 1984); if stops in Pred, further movement of the complex to T is blocked, because Pred is a functional head (20a). Verbs do ni encounter this problem, because they do not need another head to introduce the subject (20b). Further empirical consequences are explored in Baker (2008). He starts from the observation that adjectives never truly agree in person with their subjects; other categories, such as v/V or T, can agree in person. He claims that the generalization can be captured by assuming that person agreement involves more rigid licensing conditions than number or gender, the Person Licensing Condition (PLC, Baker, 2008: 143).
Many un- prefixed participles allow by-phrases with agent or causer interpretation; in the internet, Bruening found cases such as Biden's optimism undisturbed by Iraqi bombs, Toddler unfazed by lion encounter, Millions undisclosed by TX congressman, Steve Job’s birthday doesn’t go unnoticed by spammers, etc. His proposal is that adjectival participles — which are always formed in syntax — can be built over roots (37a), in which case there is no agent — this would roughly correspond to Meltzer’s decausatives — or over verbs with Voice, introducing an agent or causer (37b). The general strategy is to differentiate among classes of participles according to how many verbal projections the participial morphology subordinates; Gurer (this volume) uses the same strategy in her analysis of Turkish participles. The adjective head performs different operations at the same time; next to the obvious category change, it requires movement of the object of the root to its specifier. This object is a null operator that, after movement, creates lambde abstraction and, therefore, produces the right semantic type for the participle to modify a noun. It also existentially binds the agent or cause position, if there is one — an operation which gives equivalent results to saturation, not supressing the position.
Many un- prefixed participles allow by-phrases with agent or causer interpretation; in the internet, Bruening found cases such as Biden's optimism undisturbed by Iraqi bombs, Toddler unfazed by lion encounter, Millions undisclosed by TX congressman, Steve Job’s birthday doesn’t go unnoticed by spammers, etc. His proposal is that adjectival participles — which are always formed in syntax — can be built over roots (37a), in which case there is no agent — this would roughly correspond to Meltzer’s decausatives — or over verbs with Voice, introducing an agent or causer (37b). The general strategy is to differentiate among classes of participles according to how many verbal projections the participial morphology subordinates; Gurer (this volume) uses the same strategy in her analysis of Turkish participles. The adjective head performs different operations at the same time; next to the obvious category change, it requires movement of the object of the root to its specifier. This object is a null operator that, after movement, creates lambde abstraction and, therefore, produces the right semantic type for the participle to modify a noun. It also existentially binds the agent or cause position, if there is one — an operation which gives equivalent results to saturation, not supressing the position.
Behe Ne EEN SOAR NRE, ME PEAT CEE This is a good moment to move to a second problem: how the argument that is externalized is chosen. Notice, first, that all approaches agree that adjectives involve, necessarily, projecting an external argument — inside AP or outside it. The reasons for this are, however, different across proposals. Levin and Rappaport-Hovav (1986) and Meltzer-Asscher (2011a,b) are examples of theories where this is taken to be an idiosyncratic, though necessary, property of adjectives: that is what being an adjective means in the lexicon (similarly, Hale and Keyser, 2002 argued that adjectives always take a specifier which is provided by another category). In Bruening, as we have seen, this externalization is necessary because it creates an operator-variable structure that allows the adjective to act as a predicate, but it takes place in the syntax
Behe Ne EEN SOAR NRE, ME PEAT CEE This is a good moment to move to a second problem: how the argument that is externalized is chosen. Notice, first, that all approaches agree that adjectives involve, necessarily, projecting an external argument — inside AP or outside it. The reasons for this are, however, different across proposals. Levin and Rappaport-Hovav (1986) and Meltzer-Asscher (2011a,b) are examples of theories where this is taken to be an idiosyncratic, though necessary, property of adjectives: that is what being an adjective means in the lexicon (similarly, Hale and Keyser, 2002 argued that adjectives always take a specifier which is provided by another category). In Bruening, as we have seen, this externalization is necessary because it creates an operator-variable structure that allows the adjective to act as a predicate, but it takes place in the syntax
With respect to the unavailability of goals in ditransitive verbs like te//, Bruening assumes that such configurations are obtained through an Applicative head (Pylkkanen, 2002; Cuervo, 2003) (49). He further notes that (49) is not interpretable semantically, because the applicative head should relate two individuals, but in that configuration it relates and individual with a root. projecting te secona COMPUISOry afQuMeNnl alsO PrOQUuces al UNQraMMatlCal result. Work following Levin and Rappaport-Hovav (1986) has accepted the idea that externalization, although compulsory, is not sensitive to the theta role that the argument carries. In Bruening (2014), the principles that determine the (un) availability of an argument for externalization are purely configurational. Consider how he explains that unaccusatives, but not unergatives, can build adjectival participles. The difference is that unaccusative roots take a nominal complement, but not unergatives (48). Consequently, when the adjective is merged on top — given that in order to become a predicate, the adjective requires that a noun is externalized, the result with unaccusatives will be legitimate, as there is a noun, but not with unergatives.
With respect to the unavailability of goals in ditransitive verbs like te//, Bruening assumes that such configurations are obtained through an Applicative head (Pylkkanen, 2002; Cuervo, 2003) (49). He further notes that (49) is not interpretable semantically, because the applicative head should relate two individuals, but in that configuration it relates and individual with a root. projecting te secona COMPUISOry afQuMeNnl alsO PrOQUuces al UNQraMMatlCal result. Work following Levin and Rappaport-Hovav (1986) has accepted the idea that externalization, although compulsory, is not sensitive to the theta role that the argument carries. In Bruening (2014), the principles that determine the (un) availability of an argument for externalization are purely configurational. Consider how he explains that unaccusatives, but not unergatives, can build adjectival participles. The difference is that unaccusative roots take a nominal complement, but not unergatives (48). Consequently, when the adjective is merged on top — given that in order to become a predicate, the adjective requires that a noun is externalized, the result with unaccusatives will be legitimate, as there is a noun, but not with unergatives.
So far so good; if the following contrasts exhaustively classify the class of adjectival participles, then the picture that would emerge would be relatively clean: verbal participles, as presented in verbal passives, denote events. Adjectival participles denote states that come as a result of a change of state event or pure states. Embick (2004) attempts to codify these differences structurally. He assumes two possible verbal heads which can combine in a shell, a proposal reminiscent of Larson (1988): one specified Ag, denoting a process with a causer, and one specified Fient, which denotes a transition — perhaps more exactly, the final boundary of a transition (cf. Pifdén, 1997; Marin and McNally, 2011) — towards a state. He further assumes that the participial morphology is introduced in a head Aspect. Given this, the three kinds of participles are differentiated by which heads Aspect dominates (cf., for a similar approach to deverbal nominalizations, Alexiadou, 2001): (58a) is a verbal participle, (58b), an adjectival result participle, and (58c), a target participle. Interestingly, note that in Embick’s approach, there is no head ‘A. The adjectival interpretation is, thus, the epiphenome- non of having a stative predicate which does not include reference to a process, a suggestion which Kratzer (2000:8) explicitly makes. The reader may have already noted that this analysis implies that agents should be out in adjectival
So far so good; if the following contrasts exhaustively classify the class of adjectival participles, then the picture that would emerge would be relatively clean: verbal participles, as presented in verbal passives, denote events. Adjectival participles denote states that come as a result of a change of state event or pure states. Embick (2004) attempts to codify these differences structurally. He assumes two possible verbal heads which can combine in a shell, a proposal reminiscent of Larson (1988): one specified Ag, denoting a process with a causer, and one specified Fient, which denotes a transition — perhaps more exactly, the final boundary of a transition (cf. Pifdén, 1997; Marin and McNally, 2011) — towards a state. He further assumes that the participial morphology is introduced in a head Aspect. Given this, the three kinds of participles are differentiated by which heads Aspect dominates (cf., for a similar approach to deverbal nominalizations, Alexiadou, 2001): (58a) is a verbal participle, (58b), an adjectival result participle, and (58c), a target participle. Interestingly, note that in Embick’s approach, there is no head ‘A. The adjectival interpretation is, thus, the epiphenome- non of having a stative predicate which does not include reference to a process, a suggestion which Kratzer (2000:8) explicitly makes. The reader may have already noted that this analysis implies that agents should be out in adjectival
Related topics:
Lexical SemanticsTense and Aspect SystemsFormal SemanticsArgument StructureLexical aspect
580 California St., Suite 400
San Francisco, CA, 94104
© 2025 Academia. All rights reserved