Annotation guidelines
corpora of multiword expressions - version 1.2 (2020)
shared task on semi-supervised identification of verbal multiword expressions - edition 1.2 (2020)

---

Light verb constructions (LVC)

Light verb constructions (LVC) constitute a universal category. We retain the following key characteristics:

1. They are formed by a verb v and a (single or compound) noun n, which either directly depends on v (and possibly contains a case marker or a postposition), or is introduced by a preposition.
   In case of Hindi, the noun can be replaced by an adjective which is morphologically identical to an eventive noun. If you annotate Hindi, everywhere is this page when the noun is referred to, you should read the noun or the adjective.
   • 5.2_A_verb-noun
   • إتخذ إجراء → verb+direct object noun
             قام بزيارة → verb+prepositional-object noun
             أدى التحية العسكرية →verb+ composed noun
   • вземам решение
             държа под контрол
   • zum Einsatz kommen
             eine Rede halten
   • παίρνω μία απόφαση
             δίνω στα νεύρα
            δίνω τη χαριστική βολή
   • to give a lecture → verb + direct-object noun
            to come into bloom → verb + prepositional-object noun
            to make a high five → verb + compound noun
   • hacer una promesa
             poner en peligro → verb + prepositional-object noun
            tener dolor de cabeza → verb + compound noun
   • lan egin , aurrera egin
   • faire une présentation → verb + direct-object noun
            procéder à une analyse → verb + prepositional-object noun
            faire un faux pas → verb + compound noun
   • stupiti na snagu
             držati predavanje
   • chiamare in causa
             fare una passeggiata
   • odnieść sukces
            mieć wyrzuty sumienia
            wykonać rzut karny
   • fazer um aborto → verb + direct-object noun
             estar com fome → verb + prepositional-object noun
             fazer uma mesa redonda → verb + compound noun
   • a duce dorul
             a da divorț
             a da în clocot
             a da în fiert
   • biti v dvomih → verb + prepositional-object noun,
             imeti predavanje → verb + direct-object noun
   • дати на знање dati na znanje
             поднети жалбу podneti žalbu
2. The (single or compound) noun n is predicative and refers to an event (e.g. decision, visit) or a state (e.g. fear, courage). Predicative nouns are nouns that have semantic arguments, that is, they express predicates whose meaning is only fully specified by their semantic arguments:
   • 5.2_B_noun-event-state
   • قرار أخذ →noun refers to an event , there are 2 argument : a decider and decision
             كلمةألقى → noun refers to an event , there are 2 arguments : the talker and the speech
   • вземам решение → noun refers to an act or event
            давам съгласие → noun refers to an act or event
             имам притеснения → noun refers to a feeling or state
             имам готовност → noun refers to a feeling or state
   • eine Entscheidung treffen → noun refers to an event
            Angst haben→ noun refers to a state
   • παίρνω μία απόφαση, κάνω βόλτα → noun refers to an event
             έχω αγωνία, κάνω κουράγιο → noun refers to a state
   • to make a decision → noun refers to an event, there are 2 arguments: a decider and a choice
             to pay a visit → noun refers to an event, there are 2 arguments: a visitor and a visited place/person
            to have fear→ noun refers to a state, there are 2 arguments: somebody who is afraid and something frightening
             to have courage → noun refers to a state, there is 1 argument: the courageous person
   • dar un consejo → noun refers to an event, there are 3 arguments: an adviser, and advised person, and a theme
             tener valor → noun refers to a state, there is 1 argument: the courageous person
   • negar egin → noun refers to an act or event
            lo egin → noun refers to a state
   • donner un conseil → noun refers to an event, there are 3 arguments: an adviser, and advised person, and a theme
             avoir du courage → noun refers to a state, there is 1 argument: the courageous person
   • donijeti odluku → noun refers to an event
            imati osjećaj→ noun refers to a state
   • fare una domanda → noun refers to an event
            avere paura, avere coraggio → noun refers to a state
   • prowadzić rozmowy → the noun refers to an event
            mieć rację → the noun refers to a state
   • fazer uma prece → noun refers to an event, there are 2 arguments: the prayer and the thing she/he prays for
             ter sintomas → noun refers to a state, there are two arguments: the person having symptoms and the disease causing these symptoms
   • a lua o decizie , a face o vizită → noun refers to an event
            a avea curaj → noun refers to a state
   • biti v dvomih → noun refers to a state
             imeti predavanje → noun refers to an event
   • донети одлуку doneti odluku → the noun refers to an event
             имати право imati pravo → the noun refers to a state
3. We retain two sub-categories of verbs, which define two sub-categories of LVCs:
   • The verb v is "light" in that it contributes to the meaning of the whole only by bearing morphological features: person, number, tense, mood, as well as morphological aspect. This implies that v's syntactic subject is n's semantic argument. In this case, we annotate the construction as LVC.full.
     • 5.2_C_verb-light
     • نصيحةأسدى
              تاريخالصنع
               إستراتيجية ال وضع
     • давам изявление
              нанасям щети
     • παίρνω μία απόφαση, κάνω βόλτα , έχω αγωνία, έχω πονοκέφαλο → noun refers to a state
     • to make a presentation
              to pay a visit
              to have rights
              to have a headache
              to carry out a destruction
     • dar un paseo
              tener valor
               tener dolor de cabeza
     • faire une présentation
               faire une visite
               avoir le droit
               avoir un mal de tête
     • napraviti pogrešku
     • fare una presentazione
               fare una visita
               avere il diritto
               avere un mal di testa
     • odnieść sukces
              mieć rację
              cierpieć na anemię
     • realizar uma apresentação
               fazer uma visita
               ter um direito
               ter dor de cabeça
     • a face o prezentare
              a face o vizită
     • imeti predavanje , biti mnenja , biti v pomoč , delati razlike
     • вршити претрес vršiti pretres
               имати право imati pravo
   • The verb v is "causative" in that it indicates that the subject of v is the cause or source of the event or state expressed by n. In other words, the noun has semantic arguments expressed as non-subject elements in the sentence, and the subject of the verb brings an additional information, indicating the cause of source of the event/state. In this case, we annotate the construction as LVC.cause. These constructions are expected to be less idiomatic than other VMWEs and can be understood as complex predicates with a causal support verb.
     • 5.2_D_verb-cause
     • حربالأعلن
               حقوق أعطى
              أملأعطى
     • давам възможност
              нося късмет
     • κάνω αιδικία
               δίνω ικανοποίηση
               προκαλώ καταστροφή
     • to grant rights
              to give a headache
              to provoke a reaction
     • dar derecho
               dar vértigo
               causar un accidente
     • donner le droit
               donner le vertige
               provoquer un accident
     • dati mogućnost
     • dare il diritto
               dare le vertigini
               causare un incidente
     • to sprawia nam kłopot
              nakłada obowiązek na użytkowników
              dać prawo
              narazić na straty
              stawiać komuś cel
     • dar o direito
               dar tontura
               provocar um acidente
     • a da dureri de cap
     • dati ime nekomu , narediti konec nečemu
     • изнети мишљење izneti mišljenje
               задати главобољу zadati glavobolju

The following decision tree should be applied to decide whether a candidate should be annotated as a LVC.full, LVC.cause or none.

LVC-specific decision tree:

• Apply test LVC.0 - [N-ABS: Is the noun abstract?]
  • It is not an LVC, exit
  • Apply test LVC.1 - [N-PRED: Is the noun predicative?]
    • It is not an LVC, exit
    • Apply test LVC.2 - [V-SUBJ-N-ARG: Is the subject of the verb a semantic argument of the noun?]
      • Apply test LVC.3 - [V-LIGHT: The verb only adds meaning expressed as morphological features?]
        • It is not an LVC, exit
        • Apply test LVC.4 - [V-REDUC: Can a verbless NP-reduction refer to the same event/state?]
          • It is not an LVC, exit
          • It is an LVC.full
      • Apply test LVC.5 - [V-SUBJ-N-CAUSE: Is the subject of the verb the cause of the noun?]
        • It is not an LVC, exit
        • It is an LVC.cause

Note: test 10 [N-SEM] from the previous version of the guidelines (1.0) was considered unnecessary and has been abandoned in the current version of the guidelines.

Note: LVC tests are often hard to apply. If you hesitate at some intermediary test, continue to the next one, since the last tests of LVC.full and LVC.cause will help you reach your final decision.

Test LVC.0 - [N-ABS] Noun is abstract

Is the noun n abstract?

• continue to next test
  • 5.2_E_test-lvc0-yes
  • ... قرار ، علم ، أمل ، إجتماع
  • проблем , възможност , изявление , план
  • προτεραιότητα, θυμός,αγάπη,δυσκολία, λόγος,παρουσίαση,γέννηση
  • priority, anger, love, opinion, difficulty, speech, presentation, birth
  • paseo , derecho , ilusión , fe , duelo
  • pas , édition , discours , explication , lute
  • problem , mogućnost , ideja
  • priorità , rabbia , amore , opinione , difficultà , discorso , presentazione ,
  • kłopot , wysokość , praca , prawo , zysk
  • prioridade , festa , fé , nascimento , distinção , problema , gol
  • răspuns , prezentare
  • dvom , mnenje , ime , vloga , odločitev
  • мишљење mišljenje , претрес pretres , побуна pobuna , одлука odluka
• it is not an LVC
  • 5.2_F_test-lvc0-no
  • طاولة ، ورقة ، شخص ، يد
  • правя торта → a cake is a physical entity (not abstract)
           давам пари → money is a physical entity (not abstract)
           подавам ръка → hand is a physical entity (not abstract)
  • καρέκλα , τραπέζι , χέρι , άνθρωπος
  • chair, keyboard, hand, person
  • mesa , silla , mano , foto ,
  • aulki, teklatu, esku, pertsona
  • chaise , clavier , main , personne
  • stol , ruka , kruna
  • sedia , tastiera , mano , persona
  • złożyć kartkę → a sheet is a physical entity (not abstract)
           złożyć broń → arms is a physical entity (not abstract)
           bić pianę → foam is a physical entity (not abstract)
           wystawić fakturę → a bill is a physical entity (not abstract)
           mieć brata → a brother is a physical entity (not abstract)
  • cadeira , teclado , mão , pessoa , pedra
  • scaun , pian
  • oseba , mačka , kapa , avtomobil , roka
  • изнети јело izneti jelo → a dish is a physical entity (not abstract)

Some concrete nouns may be predicative (test LVC.1). For instance, a relational noun such as daughter is semantically incomplete without its argument: daughter of X, so daughter is predicative. However, concrete predicative nouns should not pass test LVC.0.

Some nouns may have both concrete and abstract interpretations. For instance, money is concrete when it refers to banknotes (paper money, bills): I didn't have money so I paid by credit card. However, money is abstract when referring to a conventional value used in transactions between people: He spent a lot of money in the mall. If one cannot be sure that the noun is used in its concrete interpretation, test LVC.0 passes.

Test LVC.1 - [N-PRED] Noun is predicative

Does the noun n have at least one semantic argument, implying that it is a predicative noun?

• continue to next test
  • 5.2_G_test-lvc1-yes
  • إجتماععقد → event with 2 arguments the meeting and the person that organize the meeting
            حوار أجرى→ event with 2 argument the discussion athe the person who contribute the discussion
  • поставям акцент → event, with two arguments: the agent and the object being emphasized
           имам право → property, with one semantic argument: the possessor of the property
  • einen Besuch abstatten → event, with two arguments: the visitor and the visitee
           Angst haben → property with one semantic argument: the entity having fear
           einen Blick auf etwas werfen → an event with two arguments the entity glancing and the entity glanced at
  • κάνω μία επίσκεψη → event, with two arguments: the visitor and the visitee
           έχω τη δυνατότητα → property, with one core semantic argument: the entity having the ability
           έχω μίσος → state, with two arguments: the entity being in state hate and the entity hated
           βγάζω λόγο → event, with one obligatory argument: the entity making the speech
  • pay a visit → event, with two arguments: the visitor and the visitee
           have strength → property, with one semantic argument: the entity having strength
           take a glance at something → event, with two arguments: the entity glancing and the entity glanced at
           make a contribution → event, with two arguments: the contributor and the beneficiary (notice that contribution could refer to both the event and the thing being contributed, but we always prefer the former reading when possible)
  • hacer una visita → event, with two arguments: the visitor and the visitee
           tener valor → property, with one semantic argument: the entity having courage
           echar un vistazo a algo → event, with two arguments: the entity glancing and the entity glanced at
  • bisita egin → event with two arguments: the visitor and the visitee
           itxaropena ukan → event with one single argument: the person who hopes
  • avoir du courage → state(property), with one argument: the entity having courage
  • imati osjećaj → property with one semantic argument: the entity having feeling
           otići u posjet → event, with two arguments: the visitor and the visitee
           
  • fare una visita → event, with two arguments: the visitor and the visitee
           avere forza → property, with one semantic argument: the entity having strength
           dare uno sguardo a qualcosa → event, with two arguments: the entity glancing and the entity glanced at
  • złożyć wizytę → event, with two arguments: the visitor and the visitee
           złożyć skargę → event, with two arguments: the complaining person and the one he/she complains about
           mieć prawo → state, with two arguments: the person having the right and the thing (s)he has the right to
           budzić zastrzeżenia → state, with two arguments: the person having reservations and the object of the reservations
  • ter fome → property, with one argument: the entity that is hungry
           ter idade para fazer algo → state, with one argument: the entity that is old enough
            In PT, we consider that the following classes of predicative nouns pass the test: diseases (gripe, trombose, infarto), physical sensations (fome, sede, sono), emotions (medo, paixão, nojo), cognitive entities internal to the cognizer (ideia, opinião, preocupação), characteristics (coragem, teimosia, fraqueza), relations (contato, conflito, amizade) and nouns expressing communication or speech acts (conversa, discussão, briga, conselho).
  • a face o vizită → event, with one argument: the entity that visits
            a avea curaj → property, with one semantic argument: the entity having courage
  • imeti predavanje → event, with two arguments: a lecturer and the people who are attending the lecture
  • поднети жалбу podneti žalbu → event, with two arguments: the complaining person and the one he/she complains about
            имати право imati pravo → state, with two arguments: the person having the right and the thing (s)he has the right to
           
• it is not an LVC
  • 5.2_H_test-lvc1-no
  • كتابه أحمد أعطى → the nounكتاب is a physical entity that does not pass test LVC.0, even though أحمد could be considered its semantic argument
            إعصارًا أحمدشهد → the noun إعصارًا is an event, but has no semantic arguments
  • Иван хвърли боклука → physical entity (not event/state)
  • Joe macht einen Kuchen→physical entity (not event/state), even though Joe could be considered a semantic argument
  • ο Γιάννης βγάζει τα ρούχα του → physical entity (not event/state)
  • Joe makes a cake → the noun is a physical entity that does not pass test LVC.0, even though Joe could be considered its semantic argument
           Joe experienced a tornado → the noun is an event, but has no semantic arguments
           Joe has a lot of money → the noun is abstract and Joe could be considered its semantic argument, but we consider that money (as well as other goods such as car and bananas) can exist independently of a possessor, so the possessor (owner) should not be considered as semantic argument of money
  • Ana tiene una bicicleta → noun is not abstract, so it does not pass test LVC.0
            Ana hace una foto → noun is not abstract, so it does not pass test LVC.0
  • pastela egin
  • Anna a un vélo → noun is not abstract, so it does not pass test LVC.0
            Anna affronte la tempête → noun is abstract but has no arguments
  • Ivan ima olovku → noun is not abstract, so it does not pass test LVC.0
  • Joe fa un dolce → physical entity (not event/state), even though Joe could be considered its semantic argument
           Joe ha vissuto un tornado → event, but has no semantic argument
  • przetrwać burzę → burza has no semantic erguments although it is abstract
  • quebrar a cabeça → physical entity, does not pass test LVC.0
            In PT, we consider that the following classes of abstract nouns do not pass this test: informational content that do not require agents (informações, notícias), natural phenomena (chuva, neve, tornado).
  • Joe a făcut o prăjitură → physical entity (not event/state), even though Joe could be considered its semantic argument
  • Janez ima avto → the person that has a car could be considered as a semantic argument, but the car is not an event or a state
  • преживети земљотрес preživeti zemljotres → земљотрес zemljotres has no semantic erguments although it is abstract

We only retain nouns n that have at least one semantic argument, which we define as a semantically mandatory and specific participant of the event or state expressed by the predicative noun.

Sometimes, it might be useful to consider verbs and adjectives derivationally related to the noun to reason about its semantic arguments.

Test LVC.2 - [N-SUBJ-N-ARG] Verb's subject is noun's semantic argument

Is the subject of the verb a semantic argument of the noun? In other words, is the verb linking the predicative noun to one of its semantic arguments that occurs as the subject of the verb?

• continue to next test
• 5.2_I_test-lvc2-yes
• لصديقه نصيحة أحمد قدم → أحمد is the subject of the verb and a semantic argument (Advicer ) of the noun
• Иван изнесе доклад → Иван is the subject of the verb and a semantic argument (agent) of the activity
         Президентът получи покана за посещение в Германия → Президентът is the subject of the verb and a semantic argument (the receiver) of the invitation
         Президентът получи награда → Президентът is the subject of the verb and a semantic argument (the receiver) of награда
• ο Γιάννης έκανε μία παρουσίαση στο αφεντικό του → ο Γιάννης is the subject of the verb and a semantic argument (the presenter) of the noun
• John made a presentation to his boss → John is the subject of the verb and a semantic argument (the presenter) of the noun
• María dio un paseo → María is the subject of the verb and a semantic argument (the walker) of the noun
• Max fait une promenade → Max is the subject of the verb and a semantic argument (the walker) of the noun
• Helena je otišla u posjet prijateljici → Helena is the subject of the verb and a semantic argument (the visitor) of the visit
         Susjed jedobio dozvolu za gradnju → Neighbour is the subject of the verb and a semantic argument (the receiver) of the permission
• Jan złożył wizytę Marii → Jan is the subject of the verb and a semantic argument (the visitor) of the visit
         Piotr dostał pozwolenie and budowę → Piotr is the subject of the verb and a semantic argument (the receiver) of the permission
         Beata ma marzenia o spokoju → Beata is the subject of the verb and a semantic argument (the possessor) of the dreams
         wyborcy ponoszą za to winę → wyborcy is the subject of the verb and a semantic argument (the agent) of the guilt
         ustawa budzić zastrzeżenia → ustawa is the subject of the verb and a semantic argument (the theme) of zatrzeżenia
• Felipe tomou dois banhos → Felipe is the subject of the verb and a semantic argument (the person taking a shower) of the noun
• Ion i-a făcut o prezentare șefului său → Ion is the subject of the verb and a semantic argument (the presenter) of the noun
• In Janezovo predavanje o slovenski kulturi za študente prevajalstva, the 3 syntactic arguments are expressed as a modifier with a possessive marker (Janezovo ) and prepositional phrases (o slovenski kulturi and za študente prevajalstva )
• Бранко је добио постављење Branko je dobio postavljenje → Branko is the subject of the verb and a semantic argument of the appointment (receiver)
          Јелена је Бранку узвратила посету Jelena je Branku uzvratila posetu → Jelena is the subject of the verb and a semantic argument of the visit (visitor)
• Go to test LVC.5
• 5.2_J_test-lvc2-no
• خطاب المراسل ال قاطع → المراسل that is , the subject of the verb, is not a semantic argument of خطاب , since a speech does not necessarily have an interrupter
• Приятелят на Мария прекъсна нейния доклад → Maria's friend, that is, the subject of the verb, is not a semantic argument of the report, since a report does not necessarily have an interrupter
• το αφεντικό του Γιάννη διέκοψε την παρουσίασή του → το αφεντικό του Γιάννη , that is, the subject of the verb διέκοψε, is not a semantic argument of the noun predicate παρουσίαση , since a presentation does not necessarily have an interrupter
• John's boss interrupted his presentation → John's boss, that is, the subject of the verb, is not a semantic argument of the presentation, since a presentation does not necessarily have an interrupter
          The report provides information about the economy → information only has one argument, which is its topic. The provider/source of an information is not one of its semantic arguments.
• El periodista interrumpió el discurso → The journalist, that is, the subject of the verb, is not a semantic argument of the speech, since a speech does not necessarily have an interrupter
          El informe facilita información clave
• Le journaliste a interrompu le discours → The journalist, that is, the subject of the verb, is not a semantic argument of the speech, since a speech does not necessarily have an interrupter
          Le rapport fournit des informations cruciales
• Učenici su prekinuli le predavanje → Students, that is, the subject of the verb, is not a semantic argument of the speech, since a speech does not necessarily have an interrupter
• Marek dał mi prawo wyboru → Marek is the subject of the verb and but not a semantic argument of the right (a right usually does not need to be grated)
         Incydent ten podważył zaufanie wyborców do kandydata → Incydent is the subject of the verb and but not a semantic argument of the confidence
         komisja przeprowadziła wybory → komisja is the subject of the verb but not a semantic argument of wybory , which only requires the voters and the matter of the vote
• O jornalista interrompeu a inauguração → The journalist, that is, the subject of the verb, is not a semantic argument of an inauguration, since an inauguration does not necessarily have an interrupter
          O relatório traz informações polêmicas → information only has one argument, which is its topic. The provider/source of an information is not one of its semantic arguments.
• To define a predavanje one needs to mention three participants: the presenter, the audience and the topic of the presentation. In other words, the existence of a lecture implies the existence of its arguments.
• Демонстранти су прекинули говор Demonstranti su prekinuli говор → Protesters are the subject of the verb but not a semantic argument of the speech (a speech does not necessarily have an interrupter)
          комисија је спровела гласање komisija je sprovela glasanje → комисија komisija is the subject of the verb but not a semantic argument of гласање glasanje , which only requires the voters and the matter of the vote

It is not always easy to determine if the verb's subject is an argument of the noun. You can use the former syntactic version of this test to verify your intuitions.

Test LVC.3 - [V-LIGHT] Verb with light semantics

Is v semantically light, that is, is the semantics that v adds to n restricted to: (i) what stems from its morphological features (e.g. future, plural, perfective aspect, etc.), (ii) pointing at the semantic role of n played by v's subject?

• continue to next test
• 5.2_K_test-lvc3-yes
• قرار أخذ → أخذ adds no meaning to قرار besides that of performing an activity
          معروف قدم → قدم adds no meaning to معروف besides that of performing activity
          زيارةبقام → قام adds no meaning to زيارة besides that of performing an activity
• вземам решение → вземам adds no meaning to решение besides that of performing an act
         държа реч → държа adds no meaning to реч besides that of performing an act
          поемам отговорност → поемам adds no meaning to отговорност besides that of having a property
• eine Entscheidung treffen → treffen adds no meaning to Entscheidung besides that of performing an activity
          Angst haben → haben adds no meaning to Angst besides that of having a property.
• κάνω μία βόλτα → κάνω adds no meaning to βόλτα besides that of performing an activity
         παίρνω μία απόφαση → παίρνω adds no meaning to απόφαση besides that of performing an activity
         έχω άγχος → έχω adds no meaning to άγχος besides that of having a property
         διενεργώ έλεγχο → διενεργώ is a pure syntactic operator: in any context, it only bears tense and mood and never adds any sense to the noun
         διαπράττω ένα έγκλημα → διαπράττω is a pure syntactic operator: in any context,it only bears tense and mood and never adds any sense to the noun
         
• take a walk → take adds no meaning to walk besides that of performing an activity
         make a decision → make adds no meaning to decision besides that of performing an activity
         have fear → have adds no meaning to fear besides that of having a property
         perform a check → perform is a pure syntactic operator: in any context, it only bears tense and mood and never adds any sense to the noun
         commit a crime → commit is a pure syntactic operator: in any context,it only bears tense and mood and never adds any sense to the noun
         pay a visit → the verb in its usual sense means 'to spend some money on a visit', but here it is not used in this sense and does not add any semantics to the "visiting" event
         deliver a speech → the verb in its usual sense means 'to move from one place to another', but here it is not used in this sense and does not add any semantics to the "speech" event
         undergo a surgery → undergo adds no meaning to surgery besides indicating that the subject is the patient of the surgery
• dar un paseo → dar adds no meaning to paseo besides that of performing an activity
         tomar una decisión → tomar adds no meaning to decisión besides that of performing an activity
         tener miedo → tener adds no meaning to miedo besides that of having a property
• usain egin → the verb egin adds no meaning to the noun usain besides that of performing an activity
         lo egin → the verb egin adds no meaning to the noun lo besides that of performing an activity
• ils ont du courage → have adds no meaning to courage besides that of having a property
          ils reçoivent l’ordre de partir → receive adds no meaning to order besides indicating that the subject is the recepient of the order
          il a subi une intervention chirurgicale → undergo adds no meaning to surgery besides indicating that the subject is the patient of the surgery
• imati hrabrost → imati adds no meaning to hrabrost besides that of having a property
          donijeti odluku → donijeti in its usual sense means 'to bring', but here it is not used in this sense and does not add any semantics to event
• fareuna passeggiata → fare adds no meaning to passeggiata besides that of performing an activity
         prendere una decisione → prendere adds no meaning to decisione besides that of performing an activity
          avere paura → avere adds no meaning to paura besides that of having a property
         eseguire un controllo → eseguire is a pure syntactic operator: in any context, it only bears tense and mood and never adds any sense to the noun
         commettere un crimine → commettere is a pure syntactic operator: in any context,it only bears tense and mood and never adds any sense to the noun
         fare una visita → the verb in its usual sense means 'make', but here it is not used in this sense and does not add any semantics to the "visiting" event
         fare un discorso → the verb in its usual sense means 'to make', but here it is not used in this sense and does not add any semantics to the "speech" event
• oddać hołd → oddać adds no meaning to hołd besides that of performing an activity
         wystąpić z wnioskiem → wystąpić z adds no meaning to wniosek besides that of performing an activity
• mover uma ação judicial → to move adds no meaning to lawsuit besides that of performing an activity
          apresentar uma lesão → to present adds no meaning to lesion besides that of having a property
          estar com medo → to be with adds no meaning to fear besides that of being in a state
• a avea curaj → avea adds no meaning to curaj besides that of thaving a property
          a lua o decizie → lua adds no meaning to decizie besides that of performing an activity
• Janez ima predavanje → Janez is the subject of the verb and a semantic argument of the noun (the lecturer)
• одати почаст odati počast → одати odati adds no meaning to почаст počast besides that of performing an activity
          изрећи казну izreći kaznu → изрећи izreći adds no meaning to казну kaznu besides that of performing an activity
• it is not an LVC
• 5.2_L_test-lvc3-no
• إنتباه شد → شد indicates that the attention starts
• започвам играта → започвам adds an aspectual meaning to the noun
• eine Rede beginnen → beginnen adds an aspectual meaning to the noun Rede
• ξεκινάω μία προσπάθεια → ξεκινάω adds an aspectual meaning to the noun
• to start a walk → start adds an aspectual meaning to the noun
• comenzar un discurso → comenzar adds an aspectual meaning to the noun discurso
• oinez hasi → the verb hasi adds an aspectual meaning to the noun
• donner du courage → donner indicates the source of the courage (this would not pass test LVC.2)
         donner son avis → donner adds the information that the opinion is communicated
          Ce fait attire l'attention de la justice → attirer indicates the attention starts
• početi igru → početi adds an aspectual meaning to the noun
• cominciare un ballo → cominciare adds an aspectual meaning to the noun ballo
• wymierzyć sprawiedliwość→ wymierzyć adds an aspectual meaning to sprawiedliwość
         przejść na emeryturę→ przejść adds an inchoative (change-of-state) meaning to the noun
         propozycja budzi zastrzeżenia→ budzi add an inchoative meaning to zastrzeżenia
         dopełnić obowiązku→ dopełnić adds a fulfillment meaning to obowiązek
• entrar com uma ação judicial → to enter adds an aspectual meaning to the noun
         dar uma opinião → to giveadds the meaning of communication which is not present in the name itself (one can ter uma opinião without ccommunicating it).
• a începe munca → începe adds an aspectual meaning to the noun
• Študent je prekinil njegovo predavanje → The student, that is, the subject of the verb, is not a semantic argument of the lecture, since a lecture does not necessarily have an interrupter
• отићи у пензију otići u penziju → отићи otići adds an inchoative (change-of-state) meaning to the noun
         испунити дужност ispuniti dužnost → испунити ispuniti adds a fulfillment meaning to дужност dužnost

Note that this light semantics of the verb is either usual for that verb (i.e. the verb is a pure syntactic operator, like commit, perform), or occurs in the context of the particular noun (e.g. for pay in to pay a visit). Both types of verbs pass the test.

In our view of LVCs, we do not require a light verb to be "bleached", as it is sometimes described in the literature. We simply do not take into account the relation between the verb's use as a light verb and its other uses. While the specific meaning added by light verbs to the predicative nouns have been extensively studied and described (e.g. by Miriam Butt and Tafseer Ahmed), we do not adopt any fine-grained classification here. If you have a doubt about a verb's "lightness", proceed to the next test: if you can evoke the same event/state without using the verb, then it is considered light.

Test LVC.4 - [V-REDUC] - Verb reduction

Try to build an NP without the verb, in which v's subject s becomes n's dependent. You might need to test several prepositions (of, by, for, from), possessives (my, her, somebody's), postpositions, case markers, as long as you use no verb. Can this verbless NP refer to the same event or state as the candidate v+n construction does?

• annotate as LVC.full
  • 5.2_M_test-lvc4-yes
  • دورا يلعب أحمد → دور أحمد
            تحقيق أحمد ب قام → تحقيق أحمد
  • Иван пое отговорност → отговорността на Иван — both refer to the same property/event
           Иван взе решение → решението на Иван — both refer to the same property/event
  • Paul hat eine Rede gehalten → Paul's speech both refer to the same speech event
            Ich habe ihm einen Besuch abgestattet → mein Besuch both refer to the same visiting event
  • ο Γιάννης έκανε μία παρουσίαση → η παρουσίαση του Γιάννη — both refer to the same presenting event
  • Paul had a walk → Paul's walk — both refer to the same walking event
           I paid him a visit → my visit to him — both refer to the same visiting event
           Hester gave birth to Pearl → Pearl's birth to Hester — both refer to the same birthing event (note that the key criterion is that Hester, the subject of the verb, is a (prepositional) dependent of birth in the paraphrase)
            The party gave priority to senior members → the priority of senior members for the party — both refer to the same prioritization event
  • Pedro dio un paseo → el paseo de Pedro — both refer to the same walking event
            El capitán da la orden de partir → la orden del capitán de partir
  • Pellok bisita egin zidan → Pelloren bisita -- both refer to the same visiting event
  • Paul a fait une enquête → L'enquête de Paul
            Paul procède à une perquisition → La perquisition de/par Paul
            Le général donne l'ordre de partir → l'ordre du général de partir
            Les soldats reçoivent l'ordre de partir → l'ordre aux soldats de partir
            Jean souffre de troubles psychiques → Les troubles psychiques de Jean
            Jean présente une hypersensibilité → l'hypersensibilité de Jean
            Paul reçoit des menaces de (la part de) Pierre → les menaces de Pierre à Paul
            Ce médicament présente un risque → le risque de ce médicament
           Ce fait attire l'attention de la justice → l'attention de la justice pour/sur ce fait
  • Istraživač je donio zaključak → njegov zaključak both refer to the same event
  • Paolo ha fatto una conquista→ la conquista di Paolo
            Il generale da l' ordinedi partire. → L'ordine di/da parte del generale di partire
            Paolo riceve delle minacce da (parte di) Piero → le minacce di Piero a Paolo
  • Obecni oddali hołd poległym → hołd obecnych
           Jan miał na myśli Marię → myśl Jana
           Jan otrzymał wymówienie→ wymówienie dla Jana
           Inwestycja przynosi zyski → zyski z inwestycji
  • João cometeu um deslize → o deslize do João — both refer to the same event
           O jogador cobrou um pênalti → o pênalti do jogador — both refer to the same event
           João tem consciência do perigo → a consciência do João sobre o perigo — both refer to the same state
           João recebeu a remuneração → a remuneração do João — both refer to the same event
            O paciente recebeu a visita dos familiares → a visita dos familiares ao paciente — both refer to the same event
           João apresenta lesões → as lesões do João — both refer to the same state
  • Paul a făcut o plimbare → plimbarea lui Paul — both refer to the same walking event
            i-am făcut o vizită → vizita mea — both refer to the same visiting event
  • imeti dvome → imeti adds no meaning to dvomi besides that of having a property
            delati razlike → delati in its usual sense means 'to make', but here it is not used in this sense and does not add any semantics to event
  • Професор држи предавање Profesor drži predavanje → професорово предавање profesorovo predavanje
           Овај лек представља ризик Ovaj lek predstavlja rizik → ризик од овог лека rizik od ovog leka
• it is not an LVC
  • 5.2_N_test-lvc4-no
  • في عام 2001 النور رأى أفاد التقرير بأن برنامج الصحة → نور برنامج الصحة# — one cannot remove the verb because the sense of communication is missed, so the verb is not light. As a consequence, the verbless NP ( نور برنامج الصحة ) fails to refer to the original event ( رأى برنامج الصحة النور )
  • Иван хвърли поглед на вестника → #погледът на Иван върху вестника — different semantics; and requires a different preposition
  • Paul hat einen guten Eindruck gemacht → #Paul's Eindruck auf seine Freunde has a different semantics
  • ο Παύλος πήρε νέα από τον αδερφό του → #Τα νέα του Παύλου από τον αδερφό του — one cannot remove the verb because the sense of communication is missed, so the verb is not light. As a consequence, the verbless NP (Τα νέα του Παύλου) fails to refer to the original event (Paul got news)
  • Paul got news from his brother → #Paul's news from his brother — one cannot remove the verb because the sense of communication is missed, so the verb is not light. As a consequence, the verbless NP (Paul's news) fails to refer to the original event (Paul got news)
  • Juan recibió la noticia de su hermano → #La noticia de Juan — one cannot remove the verb because the sense of communication is missed, so the verb is not light. As a consequence, the verbless NP (la noticia de Juan) fails to refer to the original event (Juan recibió una noticia)
  • Hizlariak interesa piztu zuen → #Hizlariaren interesa, #the speaker's interest -- different semantics
  • Son comportement porte une atteinte grave à l'honneur des soldats → #l'atteinte de son comportement
  • Petar je dobio poruku od direktora → #Petar's news from his boss — one cannot remove the verb because the sense of communication is missed, so the verb is not light. As a consequence, the verbless NP (Petar's message) fails to refer to the original event (Petar received message)
  • Michael Phelps pobił rekord sprzed 2 tysięcy lat→ #Michael Phelps' record
           Ulica nosi imię sławnego poety → imię ulicy
           Adam jest tego samego zdania → #zdanie Adama refers to the contents of his opinion, not to the fact of having an opinion
  • O jogador cobrou uma falta → a falta do jogador — the focus changes from taking a free kick to being one of the parts involved in a foul (it's a VID)
            O jogador provocou uma lesão → a lesão do jogador — In the reduced NP, the focus changes from hurting somebody else to getting hurt
            O músico apresenta suas composições → as composições do músico — the reduced NP does not keep the sense of presenting, it is not refer to the same event as the verbal construction
  • Paul a făcut o impresie bună → #Impresia lui Paul despre soția sa — different semantics
  • to začeti predavanje → začeti adds an aspectual meaning to the noun
  • Бранко је оборио рекорд у трци на 100 метара Branko je oborio rekord u trci na 100 metara → #Бранков рекорд #Brankov rekord

This test has a simple formulation but its application has some important subtleties which are central to our definition of the LVC.full category. The goal of this test is to keep only constructions in which the predicative noun is an event or state , excluding "gray-zone" predicates.

First, if it is not possible to build an acceptable NP where the verb v's subject s becomes a dependent of the noun n, e.g. using any preposition, postposition and/or case marker, this means that the verb is not light, and the construction cannot be annotated as LVC.full. This may remove constructions in which there is control, that is, both the noun and the verb share the same subject. However, control is not sufficient to characterize an LVC.full. In other words, LVC.4 fails, the verb is not completely light, and you cannot annotate the construction as LVC.full, even if intuitively it resembles an LVC.full due to control:

• 5.2_O_np-unacceptable
• العمل قرار أحمد أخذ → قرار أحمد بالعمل is unacceptable
• Paul a l'air de dormir → *l'air de dormir de Paul is unacceptable
          Paul a eu l'occasion de dormir → *l'occasion de Paul de dormir is unacceptable
• Zdravnik je postavil diagnozo → njegova diagnoza both refer to the same event
          Politik jedal napoved → njegova napoved both refer to the same event

Second, the fact that the NP is acceptable does not suffice to characterise an LVC.full. Furthermore, the NP version in which the verb was omitted, if acceptable, must evoke the same event or state as the LVC. Here are some tricky examples and some recommendations about how to interpret them:

• 5.2_P_test-lvc4-borderline
• جديدة اجراءت الشركة أخذت → the NP الاجراءت الجديدة is ok, the "الاجراءت " seem to refer to new procedures, so ok to annotated as LVC.full
• Имам по-голям брат → моят брат refers to one member of the relation, and not to the state of brotherhood between both actants
         отправих покана към приятелите си → покана can be interpreted both as the act of inviting and as its contents; for the first reason we count this candidate as LVC.full
• Mary has a brother → Mary's brother is a concrete NP referring to one member of the relation (does not pass LVC.0), and not to the state of brotherhood between both actants
         Mary sent a letter → Mary's letter refers to a concrete object participating in the event (does not pass LVC.0), but not to the sending event itself
          Mary has an opinion and more generally, cases of have + a noun refering to the state of having a mental content (opinion, belief) → Mary's opinion is ambiguous between the fact that she has an opinion and the content of her opinion. We recommend that these cases should be annotated as LVC.full
          Mary made a speech and more generally, cases of make + a noun refering to a speech act → Mary's speech refers to the informational content produced or communicated during the speech act, but can also refer to the act itself. We recommend that these cases should be annotated as LVC.full
          Mary made a decision → decision can refer to the deciding event (a quick decision) and/or to what is decided. We recommend that these cases should be annotated as LVC.full
• María tiene un hermano → el hermano de María is a concrete NP referring to one member of the relation (does not pass LVC.0), and not to the state of brotherhood between both actants
         María envió una carta → La carta de María refers to a concrete object participating in the event (does not pass LVC.0), but not to the sending event itself
          María dio un discurso and more generally, cases of dar + a noun refering to a speech act → el discurso de María refers to the informational content produced or communicated during the speech act, but can also refer to the act itself. We recommend that these cases should be annotated as LVC.full
          María tomó una decisión → decisión can refer to the deciding event (a quick decision) and/or to what is decided. We recommend that these cases should be annotated as LVC.full
• la compagnie a pris des mesures d'économie → the NP les mesures d'écononmie de la compagnie is ok, the semantic equivalence is difficult to judge, the "measures" seem to refer to cost-saving actions, so ok to annotated as LVC.full
• mam starszego brata → mój brat refers to one member of the relation, and not to the state of brotherhood between both actants
         Maria wysłała wiadomość → wiadomość Marii refers to the contants of the message sent by Maria, rather than to the sending event itself
         Maria jest zdania, że → zdanie Marii
• Mojca jedala Tini priložnost → #Mojčina priložnost has a different meaning; if the verb is removed, the original meaning is lost, so the verb is not light.
• Марија је послала поруку Marija je poslala poruku → Маријина порука Marijina poruka refers to the contants of the message sent by Maria, rather than to the sending event itself

Finally, some nouns, especially nominalisations, are ambiguous between events and their participants. For instance, a costruction may be an event (the construction of the bridge took 2 years) or its result (this bridge is a spectacular construction). In that case, if the verbless NP can refer to the event, then you should prefer this reading over the "participant" interpretation. For example, in John made a construction, you may ask if John's construction refers to the construction event or to its result. In this case, it can refer to the event, so it should be annotated as LVC.full.

Test LVC.5 - [V-SUBJ-N-CAUSE] Verb's subject is noun's cause

Is the subject of the verb expressing the cause of the predicate expressed by the noun? In other words, does the verb bring an additional participant to the scene, representing the source or cause of the event or state referred to by the noun?

• annotate as LVC.cause
• 5.2_Q_test-lvc5-yes
• حقوق أعطى → X has the right to Y, the granter is not a semantic argument of rights, but it causes somebody to have the right to do someting
• Иван даде възможност на Мария да представи картините си → Ivan is not a semantic argument of възможност but he is the cause of the opportunity
• to grant rights → X has the right to Y, the granter is not a semantic argument of rights, but it causes somebody to have the right to do someting
         to give a headache → X has a headache, the cause of the headache, indicated as the subject of give is not a semantic argument
         the new law provoked the destruction of the building → the destruction of X by Y, the reason for the destruction is indicated by the verb provoke, which is a prototypical causative verb. Here, the subject is not the agent of destruction, but its cause. Notice that if the sentence was the explosion provoked the destruction of the building, then the construction would be an LVC.full
          residents seek to build consensus on the development of the territory → the semantic argument of consensus is the topic on which everybody agrees, the subject of build consensus expresses an external participant responsible for the consensus to exist.
• otorgar derechos → X has the right to Y, the granter is not a semantic argument of rights, but it causes somebody to have the right to do someting
         dar dolor de cabeza → X has a headache, the cause of the headache, indicated as the subject of dar is not a semantic argument
         la nueva ley provocó la destrucción del edificio → the destruction of X by Y, the reason for the destruction is indicated by the verb provocar , which is a prototypical causative verb. Here, the subject is not the agent of destrucción , but its cause. Notice that if the sentence was la explosión provocó la destrucción del edificio , then the construction would be an LVC.full
• zadati glavobolju → X has a headache, the cause of the headache, indicated as the subject of give is not a semantic argument
• Marek dał mi prawo wyboru → Marek is not a semantic argument of prawo but he is the cause of the right
         dać podstawy prawne
         nakładać na kogoś powinność
         narazić kogoś na straty
         stawiać komuś cel
         ślady krwi wzbudziły podejrzenia policji
• Bombardamentul a provocat moartea multor civili. → Many civilians (mulți civili) died and their death (moarte) was provoked by the bombing (bombardamentul)
• Борко је Марији задао бриге Borko je Mariji zadao brige → Marija has a headache, the cause of the headache, indicated as the subject of задао zadao is not a semantic argument of бриге brige
• it is not an LVC
• 5.2_R_test-lvc5-no
• إنطباع أعطى → the subject of أعطى is not what is causing إنطباع
• Този инцидент подрони авторитета на кандидата → Инцидентът is neither a semantic argument of the authority nor its cause
• to relieve a headache → the subject of relieve is not what is causing a headache
         to give birth → tricky case, since the subject of give actually is a semantic argument of birth, so it cannot be its cause. This construction must be annotated as VID (it does not pass test VPC.4 either).
         excessive heat provokes fire → even though provoke prototypically expresses a cause, in this case fire is not predicative and should not pass test LVC.1, so the construction cannot be annotated as LVC.cause
• calmar un dolor de cabeza → the subject of calmar is not what is causing a headache
         dar a luz → tricky case, since the subject of dar actually is a semantic argument of a luz, so it cannot be its cause. This construction must be annotated as VID (it does not pass test VPC.4 either).
         un calor excesivo provoca incendios → even though provocar prototypically expresses a cause, in this case incendios is not predicative and should not pass test LVC.1, so the construction cannot be annotated as LVC.cause
• Incydent ten podważył zaufanie wyborców do kandydata → Incydent is neither a semantic argument of the confidence nor its cause (it is the opposite of the cause)
         komisja przeprowadziła wybory → komisja is neither a semantic argument of wybory not its cause
         mocny zapach uśpił czujność psów → the scent is the opposite of the cause of vigilance
• căldura excesivă provoacă incendii → even though provoca prototypically expresses a cause, in this case incendiu is not a predicate and should not pass test LVC.1, so the construction cannot be annotated as LVC.cause
• Marija ima brata → Marijin brat is a concrete NP referring to one member of the relation (does not pass LVC.0), and not to the state of brotherhood between both actants
         Marija je poslala pismo → Marijino pismo refers to a concrete object participating in the event (does not pass LVC.0), but not to the sending event itself
          Marija ima mnenje and more generally, cases of imeti + a noun refering to the state of having a mental content (mnenje, predstava, dvom ) → Marijino mnenje is ambiguous between the fact that she has an opinion and the content of her opinion. We recommend that these cases should be annotated as LVC.full
          Marija je postavila vprašanje/trditev and more generally, cases of postaviti + a noun refering to a speech act → Marijino vprašanje refers to the informational content produced or communicated during the speech act, but can also refer to the act itself. We recommend that these cases should be annotated as LVC.full
• Борко ће Марију ослободити брига Borko će Mariju osloboditi briga → Borko, the subject of ослободити osloboditi is not what is causing бриге brige

Constructions annotated as LVC.cause involve:

1. verbs that are typically used to express the cause of predicative nouns in general (e.g. cause, provoke), or
2. verbs that are only used to express the cause of particular predicative nouns (e.g. grant in to grant a right).

When the construction involves a typically causative verb (e.g. cause, provoke), it might seem counter-intuitive to annotate it as VMWE because it looks perfectly regular, not presenting any VMWE idiosyncrasy. However, it turned out difficult to distinguish idiosyncratic from regular LVC.cause, so both should be annotated, like for LVC.full. In other words, some LVC.cause constructions are compositional and can be understood as complex predicates with a causal support verb, regardless of their compositionality.

Typically causative verbs (e.g. cause, provoke) can sometimes be light. In this case, according to the LVC decision tree, LVC.full has priority over LVC.cause. For instance, the announcement provoked an unexpected reaction should be annotated as LVC.full and not LVC.cause, although provoke is a typically causative verb. Indeed, reaction has two arguments (reaction of X to Y), one of which is the subject of the verb (test LVC.2 passes). In other words, typically causative verbs may be used in either LVC.full or LVC.cause, depending upon whether the cause subject of the verb is a normal, canonical argument to the predicative noun (LVC.full) or an "external" non-canonical cause (LVC.cause).

Some verbs could be considered causative, but their interpretation goes beyond purely indicating the cause of the event/state. Therefore, you should NOT annotate as LVC.cause constructions involving:

• verbs which encode a manner of causation:
  • 5.2_S_lvc-causal-manner
  • to call a meeting entails communication to schedule the meeting
            to hold a meeting entails leadership
           to organize classes entails preparation
• verbs which encode modality:
  • 5.2_T_lvc-causal-modality
  • to allow dialogue entails permission
            to foster dialogue entails assistance
            to require dialogue entails necessity
• aspectual verbs whose subject is a semantic argument of the noun:
  • 5.2_U_lvc-causal-aspect
  • we started the meeting
            we ended the meeting
           we continued the meeting

Problematic cases and remarks

---