Adjective-Adverb Interfaces in Romance Ontology https://gams.uni-graz.at/o:aaif.ontology Zentrum für Informationsmodellierung, Universität Graz 2019 AAIF Ontology Christopher Pollin Gerlinde Schneider Katharina Gerhalter https://creativecommons.org/licenses/by/4.0 1.2 Annotation Model: Criteria for Linguistic Categorization in the Database Adjective Adverb Interfaces in Romance coordinated Coordination means the combination of two different adverbs. If two or more different adverbs are coordinated, they usually share the same syntactic function and syntactic scope. For example, two manner adverbs can be coordinated: Sp. volar alto y rápido. In those cases, both adjective-adverbs has been lemmatized and tagged and their attribution target (syntactic scope) is the same verb in the sentence: volar. For both alto and rápido, the category modifications holds "true/yes". Coordination is possible between different types of adverbs: periphrastic adverbs (prepositional phrases), adjective-adverbs and mente-adverbs. Every adverbial involved has been tagged and lemmatized with the corresponding categories. If one of the coordinated adverbs has a proper lexical form (Sp., Fr. bien, mal), it has not been tagged and only the other coordinated adjective-adverb or prepositional phrase or mente-adverb has been tagged. The category Coordination holds for coordinated verbs that may be modified by the same adverb. The adjective-adverb presto modifies both verb forms cae and sube. Therefore, both verb forms has been lemmatized and annotated regarding their syntactic construction (intransitive) and both have been tagged with "coordination = yes / true". volar alto y rápido más presto cae que sube feature attribution target Defines the attribute target of an adverb. has Contraction has Gender Defines the gender of an article, possessive or a subject. has Adverb Defines an adverb in a phrase. has Article Defines an article in a phrase. morphosyntactic structure Defines the morphosyntactic structure of an adverb. has Phrase A record in a corpus. has Preposition Defines a preposition in a phrase. has Subject Defines a subject in a phrase. has Verb Defines a verb in a phrase. inflection Defines the number of an article or subject. number semantic classification Defines the semantic classification of an adverb. syntactic construction Defines the syntactic construction of a verb. text Defines a text node. prepositional phrase annotated text Defines a text node, including the annotated words. lemma The lemmatization unifies orthographic variation –- especially regarding historical data -– and enables the search via lemma. It further allows analysis of type-token-frequencies. For the category adverb, we used the present-day, uninflected form, which is identical to the masculine singular form: Spanish claro, cierto, French droit, juste, etc. The basic lemma of derived adverbs (like Sp. ciertamente, Fr. justement) is also the adjectival lemma (Sp. cierto, Fr. juste). Therefore, the lemmatization allows for searching all adverbial variants of one lexeme. For example, the Spanish lemma seguro catches the following adverbs: seguro, segura, seguramente, de seguro, etc. For the word-class verb, we used the infinite form as a Lemma (Sp. caminar, Fr. aller). Regarding the variation of verb forms in southern Italy, we chose a reconstructed lemma marked with an asterisk and in capital letters (for example *COMMOGLIARE for the Neapolitan verb forms accumigliateva and commoglio). Also the prepositions were lemmatized (de, a, por, etc.). https://gams.uni-graz.at/o:aaif.lemma#rozzo modification of the adverb The category modification applies if the adverbial itself is modified (mostly intensified) by another adverb like Sp. muy, tan, Fr. très, trop, etc. This modifiers have not been tagged or llematized itself, since they are lexical adverbs that are not tackeled in the aaif-project. Therefore, examples like Sp. volar muy alto are tagged for the verb volar and the adjective-adverb alto, which is categorized as "modification: true/yes". volar muy alto Lemma has Overt Subject Defines a Subject as an overt subject Overt Subject The category overt subject allows to annotate null-subjects in pro-drop languages. In those cases, a placeholder has been annotated with Gender and Number and allows to show the concordance of adjective-adverbs with the null-subject as in Sp. [null-subject fem. pl.] vuelan altas, in case the reference of the null-subject is clear from the wider context of the example. On the other hand, to show the uninflected use of adjective-adverbs, possibly there is a null-subject that does not concord with the adjective-adverb: Sp. [null-subject fem. pl.] vuelan alto. vuelan altas Morphosyntactic Feature is part of text Defines a verb as part of text. Part of text The category part of text allows to annotate those verbs that are not explicitly mentioned in the annotated KWIC-Phase, but can be reconstructed from the larger context. For example, if the verb modified by an adjective-adverb has been supressed (ellipsis) to avoid repetition, a placeholder is annotated and lemmatized. For example, in the sentence Sp. María canta bien […] Pero Pedro horrible the adjective-adverb horrible modifies an elliptical verb form Pedro canta horrible. In order to be able to search for the lemma combination cantar and horrible, a verb-placeholder has been annotated with the lemma cantar and has been tagged with the category "part of text: no". María canta bien […] Pero Pedro horrible person Defines the person of a possessive. position Defines the position of a word in a phrase. reduplicated adverbs The category reduplication refers to the repetition of the same adverb. In case an adverb is reduplicated, only one form has been tagged as a single-word adverb and annotated as reduplicated. In some dialects of Southern Italy, reduplication of adjective-adverbs is especially productive and in some cases, the reduplicated form has been lexicalized with a specific meaning: Neapolitan sano sano 'entirely': ma la diceno tutta sana sana. In that case, sana has been tagged with the same categories as single-word-adverbs, that is: lemmatized as sano and annotated as an adverb with morphosyntactic structure adjective and its specific inflection features (fem. sg.). Additionally, sana has been classified as "reduplication: yes". Also lexicalized reduplications of a noun are used adverbially: Neapolitan miccia miccia 'weak'. In those cases, one of the forms has been tagged with the same categories as single-word-adverbs, that is: lemmatized as miccia and annotated as an adverb with morphosyntactic structure noun and its specific inflection features (fem.sg.). Additionally, miccia has been classified as "reduplication: yes". In other cases, adjective-adverbs are expressively reduplicated, without a specific lexicalized meaning. Therefore, the morphosyntactic structure of reduplicated adverbs is basically an adjective-adverb or a noun, and the feature of reduplication is not found under the category "morphosyntactic structure", but as a separate category / tag. ¡Paso, paso, Pármeno, no saltes! source Bibliographic reference of an entry. Adverb Every example in the corpus contains at least one annotated adverb. Given the formal, syntactic and semantic complexity of the word-class adverb, several categories were established. Adjectival Article Subject The subject is an optional category. In some examples, the subject has been annotated, if the adverb refers to the subject of the sentence. For example, in the sentence "las aves vuelan altas", the adjective-adverb altas is inflected and shows concordance with the subject "las aves". Therefore, the attribution target of this adverb is verb and subject. las aves vuelan altas Contraction Derived Audible Inflection vivre saine Entry A record in a corpus Feminine Feminine Plural las aves vuelan altas Feminine Singular con sola una palabra de tu boca Inaudible Inflection je m’en vais seule Inflection Masculine Undefined Gender Masculine Plural Neuter Neuter Plural Neuter Singular Annotatation Entity Attribution Target In the first place, the syntactic scope of the adverbial has been tagged under the category Attribution target. Adverb or adverbial Phrase Adverbs may modify other adverbs. In those cases, the Attribution target Adverb was chosen. This second adverb has not been tagged and lemmatized because there aren't any examples of an adjective-adverb modifying another adjective-adverb, but always a mente-Adverb or a lexical adverb. harto in interpreter harto duramente mal or bien Verb Target The syntactic scope may be the verb of the sentence, being the adjective-adverb a modal adverb. In those cases, the verb form modified by the adverbial has been tagged and lemmatized with the tag for Verb. correr in correr rápido Adjective Target Adverbs also modify adjectives. In those cases, the Attribution target Adjective was chosen. The adjective modified by the adverb has not been tagged. Deverbal adjectives, that is: participles, in some cases has been classified as verbs, as this allows for tagging and lemmatizing the verb form. harto in hombres harto sabios modifies the adjective sabios vestir in the sentence con vna manta blanca como andan vestidos de ordinario Verb Object Target The Attribution target Verb and Object holds for those adverbials that modify both the (transitive) verb of the sentence and the object involved. In those cases, the verb also has been annotated. los martirios los paga caros Verb Subject Target In those cases where the adverb modifies both the action described by the verb of the sentence and the subject involved (), the category Verb and Subject has been chosen as the attribution target. In those cases, the verb also has been annotated. For all corpora but the French ones, if the Subject appears in the same sentence, it also has been annotated, since the inflection of the adjective-adverb shows frequently concordance in number and gender with the Subject. elle vivait saine la mujer vive sana Sentence Target If the adverb does not modify a specific part of the sentence, but the sentence as a whole, the Attribution Target Sentence has been chosen (Sp. Verdad sea que, cierto, duele…). This also holds for discourse markers that are used as an utterance, for example, affirmation markers (Sp. Justo. Claro.) Verdad sea que, cierto, duele… Justo. Claro. Other Target Noun or syntagma without verb-reference The category Noun or syntagma without verb-reference as an attribution target englobes all types of noun phrases and prepositional phrases. con sola una palabra de tu boca justo después de tres días Semantic Classification In order to offer a rough semantic classification of the adverbials, six general categories have been established. Quantity / Intensity Adverbs that quantify or intensify a denoted quality on a scale has been tagged with the semantic classification of quantity / intensity. hombres harto sabios Manner Manner-adverbs typically describe or characterize the action expressed by the verb of the sentence, therefore most examples of manner adverbs have the verb as their syntactic scope (attribution target) and also has been tagged for the verb. correr rápido vive saine Time Adverbs expressing temporal circumstances of the verb action have been tagged with the semantic classification of time. más presto cae que sube Location Adverbs expressing local circumstances of the verb action have been tagged with the semantic classification of location. volar alto Discourse Adverbs that express metalinguistic or extra-propositional characterizations (e.g. discourse markers and sentence adverbs) has been tagged with the semantic classification of discourse. Verdad sea que, cierto, duele… and ¡Claro! Specification Focus Adverbs has been tagged with the semantic classification of specification. con sola una palabra de tu boca Undefined All undefined semantic classifications. Syntactic Construction The syntactic construction takes into account transitivity and reflexivity of the verb in every specific example annotated. Reflexive Reflexive verbs are used with a reflexive pronoun. No se vestía rico je m'en vais seule Intransitive Intransitive for those examples in which the verb does not select an object. María camina rápido Transitive As for the first feature, the category transitive holds for those sentences in which the verb is used with an explicit object. los martirios los paga caros Phrase Defines a phrase within an entry in which there are annotations. Plural Preposition Number The category Number chooses Singular, Plural and undefined. For example, "las aves in las aves vuelan altas" has been categorized as "Gender: feminine" and "Number: Plural" las aves in las aves vuelan altas Singular Undefined Number Subject The subject is an optional category. In some examples, the subject has been annotated, if the adverb refers to the subject of the sentence. For example, in the sentence "las aves vuelan altas", the adjective-adverb altas is inflected and shows concordance with the subject "las aves". Therefore, the attribution target of this adverb is verb and subject. las aves vuelan altas Uninflected Verb In some examples, the verb of the sentence has been tagged. It is, therefore, an optional category. The verb of the sentence is of interest when the adverb modifies the verb. In those cases, the verb has been lemmatized (see lemma), which allows for analysing specific verb + adverb combinations and their type-token-frequencies. As for periphrastic verb constructions, the verb carrying the main semantic information has been tagged and lemmatized: for example, Sp. vestir in the sentence con vna manta blanca como andan vestidos de ordinario. The tagged verbs has been annotated regarding three formal categories: syntactic construction, coordination and part of text. vender caro parler fort, parler bas, etc. Gender The category Gender chooses masculine, feminine, neuter or undefined. Morphosyntactic Structure The morphosyntactic structure takes into account the formal structure of the adverbial as well as its possible inflection. The annotation model suits (i) simple (one-word) adverbs (ii) reduplicated adverbs and (iii) phrasal (periphrastic) adverbs. Adjectival If the morphosyntactic structure of the adverb is a simple adjective, this means an adjective is used as an adverb, like alto in Sp. volar alto, and droit in Fr. aller droit. Therefore, adjective-adverbs are adverbs that formally use an adjective for expressing adverbial functions. The lemma of these adjective-adverbs is the uninflected form (=masc.sg.), e.g. Sp. alto and Fr. droit. If the adjective-adverb is inflected, like Sp. directos in vamos directos a casa, that information can be specified via the category inflection. The database only contains adjectives with adverbial functions, but not adjectives with adjectival syntax (e.g. el arbol es alto). The diverse adverbial functions (e.g. verb-modification as modal adverb like in volar alto) are specified under the categories attribution target and semantic classification. The database does not include lexical adverbs like Sp. / Fr. bien, mal. volar alto aller droit Noun MS The category morphosyntacic structure / noun applies to nouns that are used as adverbs. For example, in Spanish pasarlo bomba, the noun bomba is used adverbially meaning 'well'. Nouns also appear in periphrastic or phrasal adverbs. The category inflection allows for specifying number and gender of the adjective or noun used adverbially. Most Adjective-Adverbs currently are used in their uninflected form, that is, the masculine singular form: Sp. volar alto, Fr. aller droit use the masculine singular of the adjective as unmarked adverbial form. pasarlo bomba volar alto aller droit derived -mente Derived adverbs are formed by a derivational suffix like -mente that adds to an adjectival base. The category –mente covers all variants of this suffix in several Romance languages: Fr. –ment, and Sp. / Pt. / It. –mente, as well as historical variants like Sp. –mientre, -miente or –mentre. The lemma of the derived adverb is the lemma of the base adjective: for example, Sp. ciertamente can be found via the lemma search for cierto, and Fr. justement via the lemma juste. Since the mente-Adverbs take systematically the feminine singular form of the adjective, we consider that a fossilized form of the stem, but not an inflectional process on the whole adverb. Therefore, mente-adverbs are categorized as uninflected. We only consider mente-Adverbs inflected if the whole adverb takes a plural suffix, like in Sp. los ríos meramentes flotables. ciertamente justement los ríos meramentes flotables derived -eşte Romanian suffixes –eşte. derived -iş Romanian suffixes -iş. iară româneşte derived -ul Romanian suffixes –ul. dreptul derived -one/-oni The Italian suffix –oni/-one si used for actions that involve a body part touching the ground. ginocchione/i, bocconi derived -ly Romanian suffixes –ly. to run quickly Other morphosyntactic structure Other morphosyntactic Structures First Person Second Person Third Person Person Possessive In case the phrasal or periphrastic adverb includes a possessive, it is tagged for the categories Gender (masculine, feminine, neuter, undefined), number (singular, plural, undefined), and person (1st, 2nd and 3rd person). For example, the Sp. possessive mis in a mis solas has been categorized as "Number: plural, Gender: undefined, person: 1st". possessive mis in a mis solas Syntactic Category Annotatation Feature