This paper aims at analysing English structures in which a wh-moved subject triggers agreement both in the clause it is extracted from and in the immediately higher clause. This pattern is only accepted by some native speakers, and it is also attested in corpora. Although the relevant structures could at first sight be analysed as extragrammatical ‘blends’, we propose that they are in fact part of certain speakers’ linguistic competence, and hence generated by the grammar of those speakers. Adopting the approach to subject extraction developed in Rizzi & Shlonsky (2007), we suggest that extracted subjects can exceptionally be ‘hyperactive’ (Carstens 2011), and thus take part in A-relations (case and agreement) in more than one clausal domain.
We test adverb-verb word orders in Peruvian Spanish against analyses of verb movement (Pollock 1989, Embick & Noyer 2001). While the preferred order is V-adv-O, the alternative Adv-V-O is also possible. We propose that the verb raises in overt syntax and morphological insertion targets either the higher or the lower position. In the latter case, morphological requirements force the more computationally costly option of T-to-V lowering. We analyze the ungrammaticality of neg-adv-V as a blocking of the selectional restriction requirements of neg (the extended verbal projection, including T) by the intervening adverb. This distribution is parallel to English do-insertion in negative contexts (I don’t frequently eat vs. * I not frequently eat), where neg selects for a -T category (cf. Williams 1994).
The paper investigates the licensing of locative DPs and deictic adverbs in subject position in Brazilian Portuguese (henceforth, BP), taking into consideration specifically the grammatical status of third person verb inflection in this language. We develop a unified analysis for related phenomena, which allow to identifying Brazilian Portuguese (BP) as a partial null subject language (cf. Holmberg 2010). In particular, we examine (i) sentences with a locative preverbal adverb/pronoun, which can be either overt or null (further allowing a locative adverb in postverbal position), in which the verb bears third person inflection and the subject has an arbitrary reading; (ii) VS word order sentences with an overt locative adverb, giving rise to a locative inversion configuration; (iii) sentences with weather verbs and a preverbal locative DP licensing agreement on the verb; (iv) the so-called topic-subject constructions, with a (selected) DP raised from an internal VP position. We assume that third person inflection on the verb, unlike first and second person inflection, is unable to license referential definite null subjects (cf. Rabelo 2010). We claim that the possibility of filling the subject position with a locative pronoun/adverb or a locative DP is due to the fact that the third person inflection in BP is no longer referential. In turn, null or overt locative adverbs/pronouns in preverbal position can check the EPP feature in these constructions, yielding arbitrary interpretation of the external argument as a consequence of the absence of a referential feature on (third person) T. We further claim that the (null) adverb/pronoun bears a locative feature in locative inversion constructions, existential predicates, and (internal) locative/part-whole arguments.
This article shows that the pronominal system of Paraguayan Guaraní provides evidence that 3p (null) possessor pronouns behave differently from 1p and 2p (overt) possessor pronouns and argues that this difference can be captured by a conjunction of hypotheses, namely, that the 3p possessor pronoun in Paraguayan Guaraní is negatively specified for Person feature and that its syntax is crucially different from that of 1p and 2p possessor pronouns. The null possessor pronoun is an nP (not a DP) that receives its interpretation via local binding. We furthermore argue that all inalienable possessor pronouns in Paraguayan Guaraní originate as an nP and that 1p and 2p pronouns must sideward-merge with a D and then merge as specifiers of higher heads (either in the nominal or verbal domain) to be syntactically realized as full DPs. We extend the analysis proposed for inalienable possessors to a certain class of transitive verbs (known as triforme verbs) as well as to nominal possessor constructions.
This paper is a study of the adverbial modifier bien and its use as an elative or operator of extreme degree in a Caribbean Spanish variety, Puerto Rican Spanish. It is argued that there are several properties setting the elative interpretation of bien apart from other well-known uses in other varieties and from the intensifier muy. The modifier bien is also argued to be exclamative and modal in nature, related to the high degree of commitment of the speaker with the truth of a proposition.
In this paper we focus on the meaning of definite DPs that allow a weak reading. We review three different proposals for weak definites, and we present a new analysis with special reference to Romance languages. We submit that the eventual weak reading of a definite DP and its contribution to a ‘familiar’ kind of activity is exclusively dependent on whether certain stereotypical information encoded on the N present in the DP is activated at the time of utterance interpretation. These DPs do not refer to kinds and do not correspond to incorporated objects. Hence, their interpretation is not compositionally driven, but rather pragmatically inferred. We predict that the identification of weak definites takes place beyond grammar and is constrained by encyclopedic information.
Assuming that syntax and morphology constraints can target ‘situation size’ in semantics, this paper argues for the necessity of constraints on ‘big’ situations, and for their grammaticalization. Data from Bulgarian indicate that complex (viewpoint) aspectual interactions between Perfective verbs in the Imperfect Tense in adjunct/restrictor clauses and verbs in the Imperfect Tense in main/nuclear scope clauses trigger habitual interpretations. Such interactions result in propositions that can only be true in ‘big’ situations, informally described as ‘non-accidental generalizations on repeated actions that are complete’. Furthermore, a morphological contrast between Perfective Imperfects and Perfective Aorists in adverbial adjunct clauses accounts for restrictions on the modal interpretations of imperfective aspect associated with a Viewpoint operator IMPF, and distinguishes between ongoing and habitual readings.
Kennedy & Levin (2008) argue that the aspectual properties of so-called degree achievement (DA) verbs (e.g. darken) can largely be predicted from the scale structure of the adjectives to which they are derivationally related (e.g. dark). Specifically, when the adjective is evaluated on a scale that is upper closed and the standard for the adjective to truthfully apply is the upper endpoint on that scale (i.e., when the adjective is absolute; see e.g. Kennedy & McNally 2005), the corresponding DA can be either telic or atelic. In contrast, when the adjective’s scale is open and the standard is context-dependent (i.e., when the adjective is relative), the corresponding DA is atelic. In this paper, I defend, following Kearns (2007), the position that telic interpretations of DAs are not directly a function of the standards for the adjectives from which the verbs are derived. Rather, the telic interpretation simply depends on it being possible to characterize the amount of change undergone in terms of the part structure of the event described, without reference to a specific comparison class. This conclusion will emerge from reflection on how the notions of relative and absolute standards can be recast in terms of similarity- vs. rule-based classification (as proposed in McNally 2011), extended from the adjectival to the verbal domain.
In this paper I explore Multiple Wh-Movement (MWM) in European Romance, a syntactic pattern that has been regarded as impossible (all Romance languages but Romanian excluding it; cf. Escandell-Vidal 1999; RAE-ASALE 2009; Chernova 2015, among others). After reviewing some data that qualify this well-known observation, I argue that European Spanish can actually display MWM under specific discourse conditions, some of which have not been previously reported. The paper puts forward an analysis of the facts adopting Richards’ (2010)Distinctiness, a PF condition that requires for X and Y to be morphologically or featurally different (within the same domain) for them to be linearized, a solution that places the relevant parameter in the Syntax → PF wing of the grammar.
In this paper we carry out an interface investigation of subject islands in Italian and Spanish, in which subextraction in the form of focus fronting is connected with the type of Focus implemented by the subextracted PP (Corrective vs. Mirative) and the D-properties of the subextraction site. Adopting a featural approach, we claim that subjects are not ‘absolute’ islands and variation is not idiosyncratic: subextraction is possible in the two languages under examination provided specific properties are met in the relevant D-domain and depending on the information-structural properties of the subextracted PP. More precisely, subextraction of a Corrective Focus in Italian is fully accepted from [−spec] DPs, whereas Spanish needs a [−def] feature in the same context. Subextraction to obtain a mirative interpretation, on the other hand, is only (marginally) allowed from [+def, −spec] DPs: this is attributed to the compatibility between D-features and the discourse-semantics properties of a Mirative Focus.
Korean long-distance anaphor (LDA) caki is obligatorily interpreted as de se: the attitude holder should know that the reference of caki is himself/herself. However, this restriction can be obviated when the Korean direct perceptive evidential marker –te is used. In this case, caki can be used as far as the evidence holder indicated by –te knows that the reference of caki is the attitude holder himself/herself, even when the attitude holder does not know that (Lim 2012, 2014; Lim & Lee 2012; Lim & Hoe to appear, among others). In this paper we call this puzzle de se center shift, and argue that the de se center shift in Korean should be analyzed in terms of the binding relation between an operator and caki. We also present some evidence that this puzzle should not be analyzed in terms of a pragmatically motivated binding relation, such as empathic binding and/or indirect de se (such as what is argued for Chinese ziji in Wang & Pan 2014, 2015, among others). Finally, we discuss some theoretical implications of our proposal, especially regarding the interaction between –te and different types of attitude predicates.
This article addresses the well-known restriction on stage-level readings of pre-nominal predicative adjectives in Spanish. We argue that this restriction is due to a structural difference between ‘direct modifiers’ and ‘indirect modifiers’ as has been previously proposed in the literature (Demonte 1999; Cinque 2010). New evidence for such a distinction is introduced, based on the contrasting behavior of two classes of deverbal adjectives: se-inchoatives versus ‘perfective’ participial adjectives.
This paper focuses on the different properties shown by two types of verbs that surface as transitive verbs in Spanish. The article tries to demonstrate that, beside regular transitive verbs taking a DP complement (e.g. Juan leyó un libro ‘Juan read a book’), other apparent transitive verbs – pseudotransitive verbs – (e.g. Juan golpeó al prisionero ‘Juan hit the prisoner’) hide a deep ditransitive structure in which a nominal argument is conflated into an abstract predicate. Subextraction, nominalization and quantifier scope data are used to support this claim. The analysis derives the different properties exhibited by these two types of verbs from their base argument structure and shows that they are independent from other syntactic mechanisms such as differential object marking.
Although it is descriptively useful to classify lexical items into grammatical categories, this classification is not theoretically accurate and it requires a more fine-grained analysis.
This is possible by means of a model like the one presented in this article. In it, each node of a universal syntactic structure encodes an indecomposable semantic component (which can be altered by modifiers). Lexical items lexicalize different chunks of the structure. The category to which they are related and the differences between elements belonging to the same category depend on the chunk they lexicalize.
I show that prepositions, for instance, lexicalize different chunks of the structure around the node Rel(ation) (only Rel, Rel and a modifier or only a modifier of Rel). This explains the different behavior of lexical items included under the label preposition (as can be seen, for instance, in their different possibilities of combination).
This paper offers a syntactic account of Haverling’s (1994 ff.) descriptive insights regarding the formation of both stative verbs and -sco verbs expressing change in Early and Classical Latin. In particular, the formal distinction between incorporation and conflation (cfr. Haugen 2008, 2009 and Mateu 2012, i.a.) is shown to be useful when dealing with the formation of these verbs. Following Acedo-Matellán and Mateu’s (2013) formal account of Talmy’s (1991, 2000) typology of motion events, the paper also addresses the question of why aspectual resultative prefixation is a phenomenon that is expected to be found in a satellite-framed language like Latin (e.g., cfr. inarescere ‘to start becoming dry’) but not in verb-framed languages like Catalan or Spanish.
This paper aims at analysing English structures in which a wh-moved subject triggers agreement both in the clause it is extracted from and in the immediately higher clause. This pattern is only accepted by some native speakers, and it is also attested in corpora. Although the relevant structures could at first sight be analysed as extragrammatical ‘blends’, we propose that they are in fact part of certain speakers’ linguistic competence, and hence generated by the grammar of those speakers. Adopting the approach to subject extraction developed in Rizzi & Shlonsky (2007), we suggest that extracted subjects can exceptionally be ‘hyperactive’ (Carstens 2011), and thus take part in A-relations (case and agreement) in more than one clausal domain.
We test adverb-verb word orders in Peruvian Spanish against analyses of verb movement (Pollock 1989, Embick & Noyer 2001). While the preferred order is V-adv-O, the alternative Adv-V-O is also possible. We propose that the verb raises in overt syntax and morphological insertion targets either the higher or the lower position. In the latter case, morphological requirements force the more computationally costly option of T-to-V lowering. We analyze the ungrammaticality of neg-adv-V as a blocking of the selectional restriction requirements of neg (the extended verbal projection, including T) by the intervening adverb. This distribution is parallel to English do-insertion in negative contexts (I don’t frequently eat vs. * I not frequently eat), where neg selects for a -T category (cf. Williams 1994).
The paper investigates the licensing of locative DPs and deictic adverbs in subject position in Brazilian Portuguese (henceforth, BP), taking into consideration specifically the grammatical status of third person verb inflection in this language. We develop a unified analysis for related phenomena, which allow to identifying Brazilian Portuguese (BP) as a partial null subject language (cf. Holmberg 2010). In particular, we examine (i) sentences with a locative preverbal adverb/pronoun, which can be either overt or null (further allowing a locative adverb in postverbal position), in which the verb bears third person inflection and the subject has an arbitrary reading; (ii) VS word order sentences with an overt locative adverb, giving rise to a locative inversion configuration; (iii) sentences with weather verbs and a preverbal locative DP licensing agreement on the verb; (iv) the so-called topic-subject constructions, with a (selected) DP raised from an internal VP position. We assume that third person inflection on the verb, unlike first and second person inflection, is unable to license referential definite null subjects (cf. Rabelo 2010). We claim that the possibility of filling the subject position with a locative pronoun/adverb or a locative DP is due to the fact that the third person inflection in BP is no longer referential. In turn, null or overt locative adverbs/pronouns in preverbal position can check the EPP feature in these constructions, yielding arbitrary interpretation of the external argument as a consequence of the absence of a referential feature on (third person) T. We further claim that the (null) adverb/pronoun bears a locative feature in locative inversion constructions, existential predicates, and (internal) locative/part-whole arguments.
This article shows that the pronominal system of Paraguayan Guaraní provides evidence that 3p (null) possessor pronouns behave differently from 1p and 2p (overt) possessor pronouns and argues that this difference can be captured by a conjunction of hypotheses, namely, that the 3p possessor pronoun in Paraguayan Guaraní is negatively specified for Person feature and that its syntax is crucially different from that of 1p and 2p possessor pronouns. The null possessor pronoun is an nP (not a DP) that receives its interpretation via local binding. We furthermore argue that all inalienable possessor pronouns in Paraguayan Guaraní originate as an nP and that 1p and 2p pronouns must sideward-merge with a D and then merge as specifiers of higher heads (either in the nominal or verbal domain) to be syntactically realized as full DPs. We extend the analysis proposed for inalienable possessors to a certain class of transitive verbs (known as triforme verbs) as well as to nominal possessor constructions.
This paper is a study of the adverbial modifier bien and its use as an elative or operator of extreme degree in a Caribbean Spanish variety, Puerto Rican Spanish. It is argued that there are several properties setting the elative interpretation of bien apart from other well-known uses in other varieties and from the intensifier muy. The modifier bien is also argued to be exclamative and modal in nature, related to the high degree of commitment of the speaker with the truth of a proposition.
In this paper we focus on the meaning of definite DPs that allow a weak reading. We review three different proposals for weak definites, and we present a new analysis with special reference to Romance languages. We submit that the eventual weak reading of a definite DP and its contribution to a ‘familiar’ kind of activity is exclusively dependent on whether certain stereotypical information encoded on the N present in the DP is activated at the time of utterance interpretation. These DPs do not refer to kinds and do not correspond to incorporated objects. Hence, their interpretation is not compositionally driven, but rather pragmatically inferred. We predict that the identification of weak definites takes place beyond grammar and is constrained by encyclopedic information.
Assuming that syntax and morphology constraints can target ‘situation size’ in semantics, this paper argues for the necessity of constraints on ‘big’ situations, and for their grammaticalization. Data from Bulgarian indicate that complex (viewpoint) aspectual interactions between Perfective verbs in the Imperfect Tense in adjunct/restrictor clauses and verbs in the Imperfect Tense in main/nuclear scope clauses trigger habitual interpretations. Such interactions result in propositions that can only be true in ‘big’ situations, informally described as ‘non-accidental generalizations on repeated actions that are complete’. Furthermore, a morphological contrast between Perfective Imperfects and Perfective Aorists in adverbial adjunct clauses accounts for restrictions on the modal interpretations of imperfective aspect associated with a Viewpoint operator IMPF, and distinguishes between ongoing and habitual readings.
Kennedy & Levin (2008) argue that the aspectual properties of so-called degree achievement (DA) verbs (e.g. darken) can largely be predicted from the scale structure of the adjectives to which they are derivationally related (e.g. dark). Specifically, when the adjective is evaluated on a scale that is upper closed and the standard for the adjective to truthfully apply is the upper endpoint on that scale (i.e., when the adjective is absolute; see e.g. Kennedy & McNally 2005), the corresponding DA can be either telic or atelic. In contrast, when the adjective’s scale is open and the standard is context-dependent (i.e., when the adjective is relative), the corresponding DA is atelic. In this paper, I defend, following Kearns (2007), the position that telic interpretations of DAs are not directly a function of the standards for the adjectives from which the verbs are derived. Rather, the telic interpretation simply depends on it being possible to characterize the amount of change undergone in terms of the part structure of the event described, without reference to a specific comparison class. This conclusion will emerge from reflection on how the notions of relative and absolute standards can be recast in terms of similarity- vs. rule-based classification (as proposed in McNally 2011), extended from the adjectival to the verbal domain.
In this paper I explore Multiple Wh-Movement (MWM) in European Romance, a syntactic pattern that has been regarded as impossible (all Romance languages but Romanian excluding it; cf. Escandell-Vidal 1999; RAE-ASALE 2009; Chernova 2015, among others). After reviewing some data that qualify this well-known observation, I argue that European Spanish can actually display MWM under specific discourse conditions, some of which have not been previously reported. The paper puts forward an analysis of the facts adopting Richards’ (2010)Distinctiness, a PF condition that requires for X and Y to be morphologically or featurally different (within the same domain) for them to be linearized, a solution that places the relevant parameter in the Syntax → PF wing of the grammar.
In this paper we carry out an interface investigation of subject islands in Italian and Spanish, in which subextraction in the form of focus fronting is connected with the type of Focus implemented by the subextracted PP (Corrective vs. Mirative) and the D-properties of the subextraction site. Adopting a featural approach, we claim that subjects are not ‘absolute’ islands and variation is not idiosyncratic: subextraction is possible in the two languages under examination provided specific properties are met in the relevant D-domain and depending on the information-structural properties of the subextracted PP. More precisely, subextraction of a Corrective Focus in Italian is fully accepted from [−spec] DPs, whereas Spanish needs a [−def] feature in the same context. Subextraction to obtain a mirative interpretation, on the other hand, is only (marginally) allowed from [+def, −spec] DPs: this is attributed to the compatibility between D-features and the discourse-semantics properties of a Mirative Focus.
Korean long-distance anaphor (LDA) caki is obligatorily interpreted as de se: the attitude holder should know that the reference of caki is himself/herself. However, this restriction can be obviated when the Korean direct perceptive evidential marker –te is used. In this case, caki can be used as far as the evidence holder indicated by –te knows that the reference of caki is the attitude holder himself/herself, even when the attitude holder does not know that (Lim 2012, 2014; Lim & Lee 2012; Lim & Hoe to appear, among others). In this paper we call this puzzle de se center shift, and argue that the de se center shift in Korean should be analyzed in terms of the binding relation between an operator and caki. We also present some evidence that this puzzle should not be analyzed in terms of a pragmatically motivated binding relation, such as empathic binding and/or indirect de se (such as what is argued for Chinese ziji in Wang & Pan 2014, 2015, among others). Finally, we discuss some theoretical implications of our proposal, especially regarding the interaction between –te and different types of attitude predicates.
This article addresses the well-known restriction on stage-level readings of pre-nominal predicative adjectives in Spanish. We argue that this restriction is due to a structural difference between ‘direct modifiers’ and ‘indirect modifiers’ as has been previously proposed in the literature (Demonte 1999; Cinque 2010). New evidence for such a distinction is introduced, based on the contrasting behavior of two classes of deverbal adjectives: se-inchoatives versus ‘perfective’ participial adjectives.
This paper focuses on the different properties shown by two types of verbs that surface as transitive verbs in Spanish. The article tries to demonstrate that, beside regular transitive verbs taking a DP complement (e.g. Juan leyó un libro ‘Juan read a book’), other apparent transitive verbs – pseudotransitive verbs – (e.g. Juan golpeó al prisionero ‘Juan hit the prisoner’) hide a deep ditransitive structure in which a nominal argument is conflated into an abstract predicate. Subextraction, nominalization and quantifier scope data are used to support this claim. The analysis derives the different properties exhibited by these two types of verbs from their base argument structure and shows that they are independent from other syntactic mechanisms such as differential object marking.
Although it is descriptively useful to classify lexical items into grammatical categories, this classification is not theoretically accurate and it requires a more fine-grained analysis.
This is possible by means of a model like the one presented in this article. In it, each node of a universal syntactic structure encodes an indecomposable semantic component (which can be altered by modifiers). Lexical items lexicalize different chunks of the structure. The category to which they are related and the differences between elements belonging to the same category depend on the chunk they lexicalize.
I show that prepositions, for instance, lexicalize different chunks of the structure around the node Rel(ation) (only Rel, Rel and a modifier or only a modifier of Rel). This explains the different behavior of lexical items included under the label preposition (as can be seen, for instance, in their different possibilities of combination).
This paper offers a syntactic account of Haverling’s (1994 ff.) descriptive insights regarding the formation of both stative verbs and -sco verbs expressing change in Early and Classical Latin. In particular, the formal distinction between incorporation and conflation (cfr. Haugen 2008, 2009 and Mateu 2012, i.a.) is shown to be useful when dealing with the formation of these verbs. Following Acedo-Matellán and Mateu’s (2013) formal account of Talmy’s (1991, 2000) typology of motion events, the paper also addresses the question of why aspectual resultative prefixation is a phenomenon that is expected to be found in a satellite-framed language like Latin (e.g., cfr. inarescere ‘to start becoming dry’) but not in verb-framed languages like Catalan or Spanish.