On The Status of N-V Verbal Compounds in Catalan Allison Adelman Dec. 6, 2002 Senior Linguistics Thesis 1 1.0 Introduction This thesis investigates the status of verbal N-V compounds in Catalan among three different age groups of bilingual speakers of Catalan and Spanish in Catalonia. Examples of verbal compounds appear in many articles relating to Catalan morphology and have been analyzed as a productive feature of the language (Gràcia and Fullana 1999). However, this type of compound construction is unique to Catalan among Romance languages (Pena 1991:90). Given that most speakers of Catalan are bilinguals whose second language is either Spanish or French, the construction is therefore potentially unstable. It is important to look at the changes in usage of this type of compound between Catalan speakers of different generations to explore whether the productive process of N-V verbal compounding is gradually weakening as some compounds gain more opaque meanings. This study begins with a background on the syntax of Catalan verbal compounds in section 2. Specific collected examples of N-V verbal compounds are introduced and classified according to their degrees of transparency and opacity in section 3. Section 4 presents the methods of conducting the surveys used to collect speakers’ intuitions about the verbal compounds themselves. The results of the surveys are then analyzed and discussed in section 5, focusing on the effects of specific characteristics of the speakers in connection with the degrees of opacity, usage, and productivity of the compounds. This thesis would not have been possible without the help of many people to whom I am extremely grateful. First of all, I would like to thank my advisor Kari Swingle for her guidance and valuable insights. I would also like to thank Ivette Tarrida for devoting her time to distribute and collect surveys while at home in Catalonia, and for enthusiastically offering me her assistance throughout the whole process as a native speaker of Catalan. Thanks to Rachel Wright for her time spent commenting on my drafts, and her support and encouragement throughout the semester. I also am grateful to my second reader, Rosi Song, and my student reader, Julie Corder, for their helpful suggestions, and finally to Gary McDonogh for his useful advice regarding some of the glosses. 2 2.0 Background Information: Verbal Compounds Verbal compounds are verbs composed of two or more distinct words. Catalan verbal compounds may be composed of either an N-V or an ADV-V combination, as in the following examples: (1) ull - prendre eye - take to catch someone’s eye (2) mal - dir badly - say/tell to speak badly of someone Ullprendre in (1) is an example of a verbal compound formed using the noun ull and the verb prendre. Maldir in (2) is formed by combining the adverb mal and the verb dir. ADV-V verbal compounds exist in Spanish as well. For instance, Spanish has the word maldecir (from mal - decir, or badly - say/tell), which is equivalent to the Catalan word maldir and once again means to speak badly of someone. Nevertheless, most morphologists believe ADV-V verbs such as these are “quite clearly cases of reanalysis and are not synchronically productive” (Rainer and Varela 1992:134). In contrast, the process of N-V verbal compounding claimed to be productive in Catalan by Gràcia and Fullana (1999), and seen in ullprendre, does not exist in Spanish (Clements 1989:156), making its presence in Catalan all the more interesting and valuable to study.1 1 Alcoba (1988) cites several N+V verbal compounds in Spanish such as maniatar, meaning to bind someone’s hands from mano + atar (hand + tie). However, the examples support Rainer and Varela’s analysis (1992:134) that this type of Spanish compound can be classified as N+i+V, exhibiting a consistent vowel epenthesis or change to i, and including further instances of reanalyzed words in this category, such as alicortar from ala + cortar (wing + cut), which are antiquated and have generally fallen out of use. 3 2.1 Syntax Because N-V verbal compounds are unique to Catalan among Romance languages, it is interesting to look briefly at an overview of their syntactic features, especially regarding their morphological headedness and corresponding equivalent phrase structures within the context of Catalan syntax. 2.1.1 Morphological Headedness Headedness refers to the element, or head of a word, which determines the morphological and syntactic properties of the entire word. Cabré asserts that in a Catalan compound, whether nominal, verbal, or adjectival, the element to the right always dominates the combination and imposes its category over the whole word (1994:80). Catalan is indeed a generally morphologically right-headed language,2 though the headedness of its compounds can vary according to the lexical category of the compound word (Mascaró 1985:58). In contrast, the compounds of Spanish and other Romance languages are usually morphologically left-headed (Pena 1991:92), though right-headed constructions are also possible (Lang 1990: 71). The most productive type of headed compound in Catalan is formed with a noun on the left and an adjective on the right, such as the following example taken from Mascaró (1985:64): (3) barba - roig beard - red red-bearded The compound in (3) is an adjective, taking its syntactic category from the morpheme on 2 In fact, Catalan words formed as derivatives of other words through affixation are always right-headed; it is only the compounds which can sometimes be left-headed, though most types still have heads on the right. (Mascaró 1985:58). 4 the right. Similarly, verbal compounds in Catalan are right-headed, and therefore retain the lexical category of verb from their righthand element, as the following words illustrate: (4) cama - trencar leg - break to break the leg (of an animal) (5) cor - nuar-se heart - knot to have a strong feeling of distress, anxiety, or fear Trencar and nuar-se are the righthand heads of these two verbal compounds, while the lefthand element of each word modifies the verb in some way. Mascaró characterizes this type of compound by the direct relationship between the first element of the compound and an argument of the transitive verb stem (1985:68). The nouns in examples (4) and (5) are tied to their respective verbs through the same arrangement; they restrict the meaning of the simple verb, although the verb may additionally require a syntactic object. Camatrencar in (4) is an example of a transitive verbal compound, while cornuar-se in (5) is a reflexive intransitive verb. Catalan N-V verbal compounds are composed of a noun and a transitive verb; the resulting compound verb may be either transitive or intransitive. 2.1.2 Exceptionality of the N-V Verbal Compound When considering the productiveness of a certain kind of word-formation process, it is important to look at the new words within the larger context of the language. While verbal compounds in Catalan are morphologically right-headed, the language is syntactically left-headed, meaning the verb normally appears to the left of its object within the verbal phrase. 5 The following sentences, taken from Gràcia and Fullana (1999:240), are semantically equivalent, but employ two different syntactic structures, the first with a simple verb, and the second with a compound: (6) El caçador va trenc - ar la cama a l’ocell. the hunter SA break - 3sgP the leg to the bird. 3 ‘The hunter broke the bird’s leg.’ (7) El caçador va camatrenc - ar l’ocell. the hunter SA leg+break - 3sgP the bird. ‘The hunter broke the bird’s leg(s).’ In (6) the word order is typical of VO languages such as Catalan and Spanish and contains a syntactically left-headed verb phrase, while in (7) the verbal compound uniquely presents its information in OV order because the verbal morpheme appears inside the compound on the right. The syntactic left-headedness in (6), implying a VO ordering, is characteristic of all Romance languages and Catalan is no exception. The exceptionality of the N-V verbal compound is that its morphological right-headedness (or OV ordering) seems to contradict the syntactic ordering of the language. However, this difference between morphological and syntactic headedness is a common one and occurs in English as well.4 The complement-verb order within the verbal compounds is simply further evidence that they should in fact be analyzed as single lexical entities and not as strings of multiple words obeying the rules of the Catalan verb-phrase. It has been proposed that “the complement-head order of verbal compounds in Catalan is a consequence of the original Latin OV order,” since Latin was head-final with 3 3sgP signifies the 3rd person singular preterite morpheme, and SA indicates a 3rd person singular Semi- Auxiliary verbal element needed for the preterite tense, which does not have a surface realization in the gloss. 4 For example, in English it is grammatical to say either bartend or tend the bar, although in bartend the morphological head (the verb tend) is on the right, and in tend the bar, the syntactic head of the verbal phrase, tend, is on the left, displaying VO ordering. 6 respect to complements (Gràcia and Fullana 1999:243). Moll (1952) explains the forms as simply analogous to the Latin verbs that were constructed with a verbal element in the final position, and a complement of the verb on the left. Nevertheless, Gràcia and Fullana (1999:244) argue that the morphological structure of Catalan verbal compounds cannot be solely a Latinate remnant because of the continuing productivity of this class of words. They found that speakers could correctly interpret the literal meanings of new N-V verbal compounds, even when the authors had invented the compounds themselves (Gràcia and Fullana 1999:244): (8) nas+trencar nose+break ‘to break the nose (of a person)’ pell+cremar-se skin+burn ‘to burn one’s skin’ Still, it is important to note that it is likely that only the literal meanings of verbal compounds that are new to speakers of Catalan will be appropriately interpreted. Many of the compound verbs in use today are often utilized to evoke more figurative interpretations than literal ones. Furthermore, a semantic discussion of the compounds will be presented later in this paper, as well as a discussion of the productiveness of the creation of new verbal compounds in terms of their opaque or transparent meanings. Regardless of whether Catalan verbal compounds originated in Latin, their morphosyntactic structure is unique. Their morphological right-headedness is exceptional to the syntactically left-headed VO word order found in all Romance languages including Catalan, while other Romance languages such as Spanish are more consistent in their morphological and syntactic left-headedness (Pena 1991:92). If it is true that verbal compounds are in fact becoming increasingly opaque and less productive in the Catalan of bilingual speakers of Catalan and Spanish, then this may be because of 7 the compounds’ anomalous construction with respect to other types of compounds in the lexicon of either language. 2.2 Summary Verbal compounds composed of an N-V combination are found uniquely in Catalan among Romance languages. The compounds, which may be transitive or intransitive, are morphologically right-headed, and are thought to be productive, although they display an anomalous internal OV ordering with respect to the normal VO syntactic left-headedness of the language. 3.0 Data Although many of the N-V verbal compounds in Catalan are transparent and can be understood literally as the simple verb modified slightly by its complement, most have much more opaque interpretations because they have changed in significance or gained multiple meanings over time. The transparent compounds are compositional; the meaning of the parts makes up the meaning of the whole, as in the following example: (9) cor - bategar heart - beat to have the heart beat (for the heart to beat) This type of compound seen in (9) is easily interpretable based on the meanings of the two words of which it is composed. The compounds classified as opaque can be divided further into two categories: figurative and lexicalized. Figurative compounds, like transparent compounds, are compositional, but in a figurative sense: 8 (10) ull - ferir eye - hurt to hurt someone’s eyes (because of ugliness) In (10), ullferir explicitly means to hurt someone’s eyes; yet despite this the meaning is not literal because the eyes are only figuratively hurt by seeing ugliness, making this a figurative compound. Lexicalization refers to the process through which phrases, or in this case the morpheme combinations of compound words, come to be reanalyzed by speakers over time as individual and opaque units of the lexicon, composed of fewer morphemes than when they were originally created, and often shifting in significance from the original semantic value. Lexicalized compounds are completely non-compositional as demonstrated below: (11) sang - cremar-se blood - burn to become impatient The lexicalized compound in (11) has been reanalyzed into a single verb meaning to become impatient, though the words blood and burn would not themselves indicate this interpretation. The following section provides an overview and discussion of the data, roughly classifying the compounds into categories of those with transparent meanings, those with opaque meanings, both figurative and lexicalized, and those that share traits of each. 3.1 Transparent Verbal Compounds The process of productive compound word formation relies both on the native speakers’ ability to create original combinations of words, following their language’s grammatical 9 patterns, and their ability to understand novel combinations used by other speakers. The most transparent types of verbal compounds are those in which the first element of the compound, which is the complement of the verb, modifies the verb by applying it specifically to that complement. These literal, or compositional, verbal compounds are simply more explicit versions of their simple verbs. For example, trencar means to break, while camatrencar means, more specifically, to break a leg. The following are examples of transparent Catalan verbal compounds and their glosses: (12) pell - obrir-se skin - open to have one’s skin open (for one’s skin to open) (13) pell - partir-se skin - divide to have one’s skin open (for one’s skin to open) (14) cor - bategar heart - beat to have the heart beat (for the heart to beat) The noun complements pell and cor, in (12) through (14), are affected literally by the action of the verbs. The two compounds in (12) and (13) are also reflexive verbs, and all three compounds in (12) through (14) are intransitive. Each of the three compounds is also in some way related to the body and is descriptive of a process it undergoes. Some compounds have more than one possible literal interpretation: (15) peu - calcigar foot - step on to step on with one’s foot (or to step on someone’s foot)