ArticlePDF Available

The acquisition of event nominals and light verb constructions: acquiring event nominals and light verbs



In language acquisition, children assume that syntax and semantics reliably map onto each other, and they use these mappings to guide their inferences about novel word meanings: For instance, at the lexical level, nouns should name objects and verbs name events, and at the clausal level, syntactic arguments should match semantic roles. This review focuses on two cases where canonical mappings are broken—first, nouns that name event concepts (e.g., “a nap”) and second, light verb constructions that do not neatly map syntactic arguments onto semantic roles (e.g., “give a kiss”). We discuss the challenges involved in their acquisition, review evidence that suggests a close connection between them, and highlight outstanding questions.
The acquisition of event nominals and light verb constructions
In language acquisition, children assume that syntax and semantics reliably map onto each other,
and they use these mappings to guide their inferences about novel word meanings: For instance,
at the lexical level, nouns should name objects and verbs name events, and at the clausal level,
syntactic arguments should match semantic roles. This review focuses on two cases where
canonical mappings are broken––first, nouns that name event concepts (e.g., “a nap”); and second,
light verb constructions that do not neatly map syntactic arguments onto semantic roles (e.g., “give
a kiss”). We discuss the challenges involved in their acquisition, review evidence that suggests a
close connection between them, and highlight outstanding questions.
Keywords: syntax-semantics interface, word learning, event nominals, light verb constructions,
argument structure
This manuscript is published in Language and Linguistics Compass. The
published version can be requested from the authors or can be accessed
Human languages present reliable mappings between syntax and semantics. At the lexical level,
certain syntactic categories often correspond to certain semantic categories––for example, nouns
name objects, verbs events (e.g., Grimshaw, 1981; Pinker, 1984, 1989). At the clausal level,
syntactic arguments match with semantic roles––for example, “the cat is chasing the mouse”
names an event with two participants, with the cat as agent and the mouse as patient (e.g.,
Gleitman, 1990; Landau & Gleitman, 1985). These mappings have been shown to be an important
assumption children entertain in early lexical acquisition. But there are exceptions. For example,
at the lexical level, some nouns called event nominals” (e.g., “a nap”, “a party”) denote events,
rather than objects; at the clausal level, some sentences have more syntactic arguments than
semantic roles, like “light verb constructions” (e.g., “Ann gave a kiss to Joe.”). How do child
learners fare with these cases? This paper offers a review of the acquisition of event nominals and
light verb constructions, in the context of syntax-semantics mappings. We discuss existing
evidence, highlight outstanding puzzles and questions, and raise a proposal, in the hope that this
review will lay down some directions for future work regarding this topic.
In Section I, we briefly review the mounting evidence for children’s reliance on the syntax-
semantics mappings. In Section II, we discuss the acquisition challenge of event nominals, and
review recent evidence that shows light verb constructions are important cues mitigating this
challenge. Light verb constructions, however, by themselves present salient challenges; we discuss
this in Section III, at the end of which we will see a chicken-and-egg problem. In Section IV, we
propose a possible solution to the problem and discuss future directions.
I. Children rely on syntax-semantics mappings in early word learning
To learn a word, the child must map its form to its meaning. But in principle, for any given word
form, there are infinitely many possible meanings (Chomsky, 1981; Quine, 1960). The mappings
between syntax and semantics, as widely seen within and across languages, have been posited as
a foundation to guide the learner’s inference about novel word meanings.
At the lexical level, there is a strong correlation between syntactic categories like noun
and verb and semantic categories like object and event (e.g., Brown, 1957; Grimshaw, 1981;
Pinker, 1984, 1989). Upon categorizing a novel word as belonging to the syntactic category
noun, for example, the learner’s hypothesis space of possible word meanings can be reduced
down to objects and entities, saving the effort to consider event concepts. In past decades, many
studies have demonstrated that young learners are indeed able to use such mappings to infer novel
word meanings––for a novel noun, they hypothesize an object/entity meaning (e.g., Echols &
Marti, 2004; Fennell, 2006; Prudent, Hirsh-Pasek, Golinkoff, & Hennon, 2006; Trehub &
Shenfield, 2007; Waxman & Booth, 2001); and for a novel verb, they hypothesize an event/state
meaning (e.g., Bernal, Lidz, Millotte, & Christophe, 2007; de Carvalho, Dautriche, Lin, &
Christophe, 2017; Echols & Marti, 2004; He & Lidz, 2017; Kobayashi, Mugitani, & Amano, 2006;
Oshima-Takane, Ariyama, Kobayashi, Katerelos, & Poulin-Dubois, 2011; Oshima-Takane, Satin,
& Tint, 2008; Waxman, Lidz, Braun, & Lavin, 2009). To make such inferences, the child needs to
recognize the syntactic category (e.g., noun, verb) of the novel word; or, in other words, to
distinguish one category from another
. It has been suggested that child learners use distributional
cues in the linguistic input to do so. For example, analyses of child-directed speech have identified
frequent frames like “the__is” and “you__it” as good cues to the noun and verb categories,
The child does not need to know the category labels “noun” and “verb”. They only need to be able to distinguish
one category from another (Pinker, 1984).
respectively (Mintz, 2003).
Young learners have also been shown to be sensitive to these cues
(Santelmann & Jusczyk, 1998), and able to use these cues for novel word categorization (Bernal,
DehaeneLambertz, Millotte, & Christophe, 2010; Hicks, 2006; Höhle, Weissenborn, Kiefer,
Schulz, & Schmitz, 2004; Jusczyk & Aslin, 1995; Mintz, 2006; see Shi, 2014 for a review).
At the clausal level, semantic roles such as agents and patients tend to be realized as
syntactic functions such as subjects and objects (e.g., Dowty, 1989, 1991; Grimshaw, 1981;
Macnamara, 1982; Pinker, 1984, 1989); and verbs with n syntactic arguments tend to have n
semantic roles taking part in the events (e.g., Fisher, 1996; Gleitman, 1990; Landau & Gleitman,
1985). Inferences about the meaning of a clause/predicate could be drawn based on these
correlations, and plenty of evidence for child learners’ ability to do so has accrued. For example,
“Big Bird is tickling Cookie Monster” is taken to name an event where Big Bird is the agent and
Cookie Monster is the patient, rather than an event with reversed roles (Hirsh-Pasek, Golinkoff,
Fletcher, DeGaspe-Beaubien, & Cauley, 1985); and given a scene of a bunny sitting on the back
of a monkey, the novel verb in the clause “the money is ziking the bunny” is interpreted as
CARRYING, but that in “the bunny is zilking the monkey” is interpreted as RIDING, in absence
of richer discourse context (Fisher, Hall, Rakowitz, & Gleitman, 1994). These data all suggest
young learners use the syntactic position of a noun phrase to infer its semantic role. Learners also
use the number of noun phrases to make inferences about the kind of event denoted by the clause.
For example, “the mouse daxed the bird” is taken to describe a two-role causative event, such as
pushing, as opposed to a one-role event, such as jumping, or a three-role event, such as offering
something to someone (Arunachalam & Waxman, 2010; Fisher et al., 1994; Naigles, 1990; Naigles
To utilize these frames, children do not need to have acquired the function morphemes; they just need to recognize
them as frequently appearing units and use them as distributional cues.
& Kako, 1993; Yuan & Fisher, 2009; Yuan, Fisher, & Snedeker, 2012 inter alia). Importantly,
children still rely on the mappings even when their language allows argument drop (Göksun,
Küntay, & Naigles, 2008 on Turkish; Lee & Naigles, 2008 on Mandarin; Lidz, Gleitman, &
Gleitman, 2003 on Kannada; Matsuo, Kita, Shinya, Wood, & Naigles, 2012 on Japanese), and
even when their language offers more reliable cues than the number of syntactic arguments (e.g.,
in Kannada, morphological marking reliably cues causativity whereas transitivity does not; see
Lidz et al., 2003). Children’s bias to rely on the syntax-semantics mappings in language
acquisition, even if it is not perfectly reliable, can be taken as indicative of how deeply ingrained
this mechanism is in learners’ brains.
Debate is still ongoing about how the mappings are manifested in the learner’s mind. In
some theories, they are built into the child’s learning apparatus, together with the relevant category
notions (e.g., noun/verb, object/event, subject/object, agent/patient). According to this theory,
what the child learner needs to learn from the input is the surface realizations of those categories
in the target language (Bloom, 1990; Borer & Wexler, 1987; Pinker, 1984; Valian, 1991; Wexler,
1998). Alternatively, the mappings may be induced by the learners based on patterns in the input,
and the relevant categories are distributionally defined clusters (e.g., Ambridge, Pine, Rowland,
Freudenthal, & Chang, 2014; Lieven, Behrens, Speares, & Tomasello, 2003; Tomasello, 2000,
2003). One direction researchers pursue to distinguish these possibilities is to see whether such
mappings and relevant categories are utilized early in development, before the learner receives
enough exemplars from the input as basis for induction (see He & Lidz, 2017, pp. 1718, for a
recent discussion). But it still remains unclear how early is early enough, and how many exemplars
are sufficient. Thus far no conclusive evidence has yet appeared to resolve the debate. Regardless
of their origin, though, it appears uncontroversial that reliable syntax-semantics mappings function
as a strong assumption, kick-starting the daunting task of word learning.
II. The challenge of event nominals: Light verb constructions as a recue
The mappings discussed above only capture strong correlations between form and meaning, but
are by no means deterministic or exclusive. For instance, the correspondence between subjects and
agents only seems to hold in what Keenan (1976) calls “basic sentences” (e.g., active sentences
are basic; passives are not). For another instance, instruments are critical participants in many
events (e.g., knives in slicing events), yet are not reliably expressed as syntactic arguments (e.g.,
Rissman, Rawlins, & Landau, 2015; also see He, Wellwood, Lidz, & Williams, 2013; Wellwood,
He, Lidz, & Williams, 2015 for discussion on event participant conceptualization). Since children
heavily rely on syntax-semantics mappings in early word learning, as discussed in Section I, cases
where the mappings are disrupted would presumably pose a challenge. The case of interest here is
“event nominals”.
Event nominals are words that have nominal syntax but semantically denote event rather
than object concepts––for example, “an earthquake”, “the destruction”, “a nap”, “the birthday
party”, etc.
Despite the mismatch, this category does appear in children’s early productive
vocabulary. In fact, the MacArthur-Bates Communicative Development Inventories––Words and
Sentences (Fenson et al., 1993), a vocabulary inventory for children aged 16-30 months, includes
several event nominals like nap, party, game, or bath. In a study of 45 children, Nelson,
Hampson, & Shaw (1993) reported more than half of the children used event-denoting nouns by
In this paper, we broadly define event to generally denote things that happen over time, including states, facts, and
situations, since nothing in this paper hinges on the exact definition of event (Casati & Varzi, 2008).
20 months of age.
Similarly, in a study of 9 English-learning 1- and 2-year-olds’ productions,
Barner (2001) found that, although the majority of children’ words respected the canonical
mappings (that is, object-denoting words used in noun syntax and event-denoting words in verb
syntax), children also demonstrated flexibility, producing some object-denoting verbs and event-
denoting nouns.
But production does not entail adult-like representation; in other words, the fact that
children have event nominals in their productive vocabulary does not necessarily entail that they
know what these words mean, like adults do. In fact, a recent study with 2- to 3-year-olds
demonstrates that children’s event nominals are closely associated with the objects involved in the
event (Arunachalam & He, 2018). Children were asked to match an event nominal like nap to a
picture. When the selection was between a picture of sleeping-in-a-bed (i.e., sleeping event
happening with a canonically associated object, a bed) and a picture of a bed alone (i.e., no
occurrence of sleeping), children were significantly above chance, with a clear preference of the
event over the object; however, they were at chance when the selection was between a picture of
sleeping (not in a bed) and a picture of a bed alone. This suggests that at an age when children
already produce some event nominals, their representations are not completely adult-like, in that
they incorrectly require canonical objects to be involved. Nevertheless, by this age, they do seem
to know that words like nap are eventive.
How do children come to learn that certain nouns have eventive denotations, especially
provided that they rely on the canonical mappings such as nounobject and verbevent (see
Section I; e.g., Waxman et al., 2009, for nouns; He & Lidz, 2017, for verbs)? One possibility is
We use the word event to refer to both actions and non-action events (see footnote 3). In Nelson et al. (1993), some
children produced nouns that denote actions, some produced nouns that denote non-action events.
that they first learn the verbal form of the word (e.g., napped), knowing that this word has eventive
meaning, and then “transfer” this meaning to its nominal form (e.g., a nap). While this may well
be the case for some event nominals, many early-acquired event nominals rarely occur in the verb
form in the language input to young children. For example, a preliminary corpus analysis of 30
CHILDES corpora of children ages 5 and under (He & Arunachalam, under review) (He &
Arunachalam, under review) reveals that nap occurs in its verb form only 9 out of 498 times,
and party 2 out of 484.
Another possibility is that children know verbs appear with arguments, based on which
they infer that nouns appearing with arguments (e.g., “the early planting of the seeds”, “the
destruction of the city”) resemble verbs in that they both have an event denotation. But first, early
acquired relational nouns (sister, father, cousin) also require arguments; and second, many early-
acquired event nominals seldom appear with arguments (e.g., ?the nap of the baby, *a jump of the
boy; Borer, 1999; Grimshaw, 1990; Moulton, 2014; Roy & Soare, 2013).
Although lacking extensive argument structure cues, there may still be other cues in the
linguistic environment of simple event nominals that signal their distinctive semantic status,
differing from regular object-denoting nouns. One such cue, as recent evidence suggests which we
review below, is occurrence in light verb constructions.
Light verb constructions are complex verb phrases (VPs) in which the verb makes the bulk
of the syntactic contribution, providing tense, number, agreement and aspect marking,
semantically the verb is partially or fully devoid of its default meaning and thus seems to be
“bleached” (Brugman, 2001; Butt, 2003, 2010; Jackendoff, 1974; Jespersen, 1954; Wiese, 2006)
But see Barner, Wagner, & Snedeker (2008) and Wittenberg & Levy (2017) for discussion how the verbal markers
semantically interact with the syntax and semantics of the event nominal.
(though not completely empty; see e.g., Everaert & Hollebrandse, 1995). This semantically-light
nature of light verbs has led some theories to consider them as functional elements such as
auxiliaries (e.g., Cattell, 1984), or a middle category between auxiliary and main verb (e.g., “thin
verbs”, Allerton, 2002). Compare (1a) to (1b) below. The verb “give” in (1a) is a regular verb,
describing a transfer event, which is the semantic core of the event. But in (1b), “give” seems to
make little contribution to the meaning of the event, yet the noun phrase “a kiss” does a lot; in fact,
the meaning of (1b) can be largely preserved in a paraphrase without the verb “give”, such as (1c)
(Wittenberg, Khan, & Snedeker, 2017; Wittenberg & Snedeker, 2014) (but see Wittenberg &
Levy, 2017, for evidence of meaning differences between (1b) and (1c)). The verb in (1b),
therefore, is a light verb, and constructions like (1b) are light verb constructions. Other examples
of light verb constructions include “take a nap”, “have a party”, “do a jump”, etc. The form-
identical counterpart in (1a) is a non-light construction. Other examples include “take a gift”, “have
a pet”, “make a knot”, etc.
(1) a. Alan gave Elsa a book.
b. Alan gave Elsa a kiss.
c. Alan kissed Elsa.
Recent evidence suggests that light verb constructions may be an important to cue to the
eventive denotations of event nominals. In He & Arunachalam (under review)’s corpus analysis
of CHILDES mentioned above, they find that event nominals occur with a small set of light verbs
–e.g., “nap” mostly occurs with “take” and “have”, whereas object-denoting nouns occur with a
Note that light verb constructions are hallmarked by the verb’s little semantic contribution, and are not limited to a
particular argument structure. In other words, a light verb construction could be transitive (“take a nap”) or
ditransitive (“give Elsa a kiss”). We therefore include a variety of examples here.
much more diverse set of verbs––e.g., “cake” occurs with “eat”, “buy”, “bring”, “have, a subset
of which have the same surface forms as light verbs but are not semantically light (as in have
some cake”). In other words, in the input available to child learners, object nouns occur in a wide
range of verbal contexts, whereas eventive nouns occur always with a small set of verbs with
vague, basic meanings. This distributional difference could be a first cue for the child to separate
event nominals from regular object-denoting nouns.
Stronger and more direct evidence comes from a novel-word learning study (Arunachalam
& He, 2018; He & Arunachalam, under review). Preschoolers (2- to 5-year-olds) first learned a
novel noun from a dialog between two conversers––one group of them heard the word in light
verb constructions (e.g., “I can do a gorp”, “Annie likes doing gorps”), and the other group heard
it in non-light syntax (e.g., “I like gorps”, “I found a gorp.”); both groups then saw a novel
character performing a novel action, and heard the novel noun again (e.g., “Look, he’s doing a
gorp!” for the light-syntax group, and “Look, there’s a gorp!” for the non-light group).
Subsequently at test, the children saw two scenes on opposite sides of the screen, both depicting
the familiar character––in one it performs a new action, and in the other it performs the now-
familiar action––and were prompted to point to the novel noun’s referent (e.g., “Can you point to
a gorp?”). For children who heard the non-light syntax, they should infer that the novel noun
labeled the now-familiar object, based on the canonical mapping between nouns and objects. For
children who heard the light syntax, however, if they were able to use light verb constructions as
a cue to override the canonical mapping, they should infer the novel noun labeled the novel action.
The results were clear: Children who learned the novel noun in light verb constructions pointed to
the familiar action (the target) 100% of the time, whereas children who learned it in non-light
syntax pointed to each character randomly (58% point to the target). No age effect was found––
children at all ages (within the relatively wide age range) successfully utilized light verb
constructions to infer an eventive denotation. But note that the only light verb used in this study
was “do”, which is special among all light verbs––Butt (2010) calls it a “verbalizer” rather than a
typical light verb. We will return to this point later.
In sum, in this section, we see that event nominals present a case of mismatch between
syntactic and semantic categories, posing a potential learning challenge; yet light verb
constructions provide useful cues that a noun names an event, rather than an object, concept. But
we are yet faced with another puzzle: how do children acquire light verb constructions to begin
III. How children acquire light verb constructions: Challenges and proposed solutions
Children’s success in acquiring event nominals via light verb constructions is impressive, but
puzzling. After all, light verb constructions are a rather complex construction by themselves. In
this section, we will discuss (i) light verb constructions in children’s production, and (ii) challenges
in recognizing light verb constructions (ii1 & ii2).
(i) Late production of light verb constructions
While the verbs used in light verb constructions themselves (do, give, make, take) are
acquired early (Goldberg, 2013; Maouene, Laakso, & Smith, 2011), true light verb constructions
appear late in children’s spontaneous production (Barner, 2001). In Barner's (2001) study with 9
English-learning children (aged 1;3-4;6), early uses of the forms of some potential light verbs (i.e.,
verbs in identical forms with light verbs, but not necessarily light) were observed, at Brown’s Stage
I, with a 1.0-2.0 MLU; but these early uses were hardly light verb constructions––only half took
complements at all, among which the majority occurred with pronouns (e.g., “do it”, “get that”).
Frequent use of these verbs with content nouns as complements did not begin until Brown’s Stage
IV (MLU 3.0-4.0, roughly 40-46 months). Even when children use these verbs with content nouns
as complements, a lot of their uses are the so-called general all-purpose verbs (“GAP” verbs)––
which are used in place of verbs with specific meanings (e.g., use make a house” instead of
“build”, “make a hole” instead of “drill”)––to be distinguished from true light verbs (see a
discussion of this distinction in detail in Kambanaros & Grohmann, 2015).
The late-emergence of true light verb construction use is perhaps due to the non-canonical
mapping of this construction, and children may prefer to use a different way to express the same
meaning. And in fact, children prefer to use non-light (e.g., “A kissed B”) over light constructions
(e.g., “A gave B a kiss”) (e.g., Barner, 2001; Oshima-Takane, Barner, Elsabbagh, & Guerriero,
Similarly, in a count for this paper, we found that the same verb is used by children in non-
light more often than in light verb constructions: A perfunctory analysis of the Brown corpus
(MacWhinney, 2000) shows that of about 3,500 children’s utterances of high-frequency verbs
(do, make, have, give, get) occurring with full NP objects (as opposed to pronouns),
only 15.6% of these NPs denoted events and could thus conceivably be parts of light verb
constructions, with little variance between verbs and children’s ages (see Figure 1). This
preference of children holds despite that in adult speech, verbs like give are used more in light
than in non-light verb constructions (Piñango, Mack, & Jackendoff, 2006; Wittenberg, Paczynski,
Wiese, & Kuperberg, 2014; Wittenberg & Piñango, 2011). Together, current evidence seems to
suggest that light verb constructions appear late in children’s production; in fact, some argue that
it is the complexity of light verb constructions that delays children’s production of event nominals
(Barner, 2001; Oshima-Takane et al., 2001).
(ii) Recognition of light verb constructions: Challenges and proposals
Late production does not necessarily entail late comprehension, just as early production does not
entail early adult-like comprehension (as in the case of event nominals). Children may still be able
to recognize light verb constructions early on, and utilize them as a cue to event nominals, as
shown in Arunachalam & He (2018). In fact, early recognition of light verb constructions is likely
because light verbs are quite frequent in the input (De Villiers, 1985; Naigles & Hoff-Ginsberg,
1998; Theakston, Lieven, Pine, & Rowland, 2004), and they are semantically general––
semantically general verbs have been theorized to have a privilege in acquisition by a number of
researchers (Clark, 1978; Gerken, 1991; Goldberg, 1998; Ninio, 1999; Pinker, 1989). But there
are challenges.
(ii-1) How to distinguish a light verb from its non-light counterpart
One salient challenge in the recognition of light verb constructions is that light verbs are
often form-identical to a non-light verb of the language (at least when we exclude verbalizers such
as “do” and modal-like light verbs; Butt & Lahiri, 2013); in other words, at least in Germanic
languages, a light verb use and a non-light verb use are indistinguishable: See “give” in (1a) and
(1b) above, “took” in (2a) and (2b), and “had” in (3a) and (3b) below. If the child learner uses
light verb constructions as a cue to event nominals, and she mistakes a non-light verb for its light
counterpart, she will wrongly infer, taking (2a) and (2b) as an example, “pillow” is event denoting
(provided that she does not know the meaning of pillow); and similarly, if she mistakes a light
verb for its non-light counterpart, she will wrongly infer that nap is object denoting (provided
that she does not know the meaning of nap). How, then, does the child learner distinguish a light
verb construction from its non-light counterpart, when not all parts of the construction are in the
child’s lexicon yet?
(2) a. Al took a nap.
b. Al took a pillow.
(3) a. Jo had a walk.
b. Jo had a kitty.
Here we propose a possible solution: Light and non-light verb constructions have different
syntax-semantics mappings, which can be a cue to differentiate them. Specifically, syntactic
arguments and semantic roles match in a one-to-one manner in non-light constructions, but not in
light verb constructions. For example, both (1a) and (1b) have three syntactic arguments––“Alan”,
“Elsa”, and “a book” / “a kiss”. In (1a), each is mapped onto a semantic role in the event, a giver,
a recipient, and a theme. But the event of (1b) only seems to have two semantic roles, a kisser, and
a kissee, rendering the NP “a kiss” with no semantic role to map onto. Alternatively for (1b), one
may argue that the event has two sets of semantic roles: One set projected by the light verb,
Cases with “have” are interesting because they provide an additional cue for the child, owing to their aspectual
properties. For instance, the progressive is only possible for combinations with eventive or consumption nouns (“Jo
is having a walk/*Jo is having a kitty”, with the idiomatic “Jo is having a baby” being, as far as we can see, the only
exception that does not involve a coercion to ingestion: Jo is having a beer/burger/ice cream/soup” means she is
consuming these objects, which would, sadly, be the only available interpretation for “Jo is having a kitty” as well.)
Likewise, the simple present acquires a habitual/stative reading for the eventive noun: *“Jo has a walk” is odd; but
“Jo has a kitty” is fine, since it is a state. In short, the distribution of grammatical aspect may clue kids in to when they
are dealing with a light verb construction as well. Of course, other uses of “have” with abstract concepts denote states
relating the subject to a property (“Jo has a kidney infection”, “Jo has a fun attitude”). How do children acquire these?
We hope someone will ask this question in greater detail than we can here.
including a giver, a recipient, and a theme, in which the NP “a kiss” is the theme of transfer; and
the other set projected by the event nominal, including a kisser and a kissed; this is called
“Argument Sharing” in some theories (Baker, 1989; Butt, 2010; Jackendoff, 1974, 2002; Müller,
2016; Ramchand, 2013). In either case, the syntax-semantics mapping is not straightforwardly
homomorphic for light-verb constructions, but is so for their non-light counterparts. See Figure 2.
Recent evidence suggest that adults do distinguish light from non-light constructions in
terms of syntax-semantics mappings. In Wittenberg et al. (2017) and Wittenberg & Snedeker
(2014), adults were trained to categorize sentences based on the number of semantic roles
conveyed. They treated non-light transitive verbs (e.g, “Grandma Kennison grew marijuana
plants”) as belonging to one category (i.e. 2 roles/arguments) and sentences with non-light
ditransitive verbs (“The millionaire sells his friend a yacht”) as belonging to another category (i.e.
3 roles/arguments); yet sentences with light ditransitive verbs (e.g., “The teenage gave a kick to
his rival”) were treated as a category in between (i.e. 2 or 3). But whether children are able to do
so remains unclear.
(ii-2) How to recognize a “potential” light form
There is yet another problem with the above proposal: Not all structures that demonstrate
misalignment in argument-role mapping are light verb constructions, both within English and
across languages. For example, control structures like in (4) present similar mapping problems to
learners (Kirby, 2009; example from there), because the object of the matrix clause in these
structures is, at the same time, the subject of the controlled clause:
(4) Suki asked/told Neili [PROi to kiss Louise]
Nevertheless, children comprehend structures like (4) above by age four, and they also
produce them. A different story are subject-to-subject raising constructions like (5), whose
acquisition has been argued to be delayed until age seven (Hirsch, 2011; example from there):
(5) Johni seems (to Mary) [ti to be dancing]
But even constructions that don’t involve lexical items being assigned multiple thematic
roles from different predicates similar to light verb constructions, which assign thematic roles
from the light verb, and the light noun pose problems to learners. For instance, (6a) has two
syntactic arguments, but a stealing event necessarily entails a thief, a loot, and a victim (without
the victim, steal would equate pick up). For another example, (6b) has two syntactic
arguments, but an additional instrument role seems to be entailed in a wiping event. One may argue
that the instrument role is not a core event participant (Rissman et al., 2015), but in many languages
(e.g., Mandarin, Igbo), the instrument is realized as a syntactic argument whereas the patient is
not, as in (6c), when the patient role is almost the least controversial candidate for a core event
participant (see Williams, 2008, 2015 for more discussion).
(6) a. Mo stole a purse.
b. Al wiped the table clean.
c. Ta ca zang le mabu.
He wipe dirty LE (aspect) rag
He wiped something with a rag, and as a result, the rag became dirty.
Therefore, while a syntax-semantics mismatch may be a useful cue to differentiate a light
verb from its non-light counterpart, it is only helpful when the child knows that it is “potentially”
light (as opposed to a heavy verb with no light counterpart). How does the child know whether a
verb is a potential light verb (see (7) for our definition) in the first place?
(7) Potential light verb: A verb that can appear in light verb constructions but sometimes
does not. A potential light verb has a light verb usage and a form-identical non-light verb
usage. Examples: “take” and “have” are potential light verbs, whereas “wipe” and “steal”
are not.
One possibility lies in the semantics. Semantically heavier verbs often have more specific
restrictions on what arguments they take, therefore selecting a narrower range of object nouns,
whereas semantically lighter verbs may occur with a broader range of object nouns (e.g., Brown,
2008). For example, Brown (2008) has found that in Tzeltal, argument ellipsis usually happens
with semantically heavier verbs, likely because these verbs select for specific objects, dropping
which will not induce ambiguity (also see Resnik (1996) for evidence for correlation between verb
selectivity and object ellipsis). Maouene, Laakso, & Smith (2011) also found that adult speakers
came up with a narrower range of object nouns associated with semantically heavier verbs than
with semantically lighter ones. This difference could be a cue for deciding whether a verb is a
potential light verb.
However, two problems limit its potential. First, note that “semantically lighter” is a
relative notion––while light verbs are indeed semantically lighter, semantically lighter verbs are
not necessarily light verbs. For example, the verb help can be associated with a relatively wide
range of objects, and thus make a semantically lighter verb according to the criteria above,
compared to the verb read which typically selects for book (among others) (Maouene et al.,
2011); but help is not a light verb. If the child learner uses how many semantically-associated
objects as a cue, she may miscategorize verbs like help as a potential light verb. In addition,
exactly because this is a relative notion and semantically lighter and heavier verbs are distributed
on a continuum, rather than in two discrete clusters (Maouene et al., 2011), there is no category of
“light verbs” that the child learner can arrive at based on this cue. Second, and more importantly,
although a distinction between semantically lighter and heavier verbs in terms of the range of
objects associated is found in adult speakers, no such distinction is found in the language in
children’s environment; sampling from CHILDES data
suggest that most verbs seem to all co-
occur with a narrow set of nouns. For example, push is associated with a quite wide range of
objects according to adult speakers, but in CHILDES, “push buttons” accounts for the majority of
occurrences of push (Maouene et al., 2011). Thus, this cue may not be quite useful to children
after all.
Here we propose another possibility: The child infers that a verb is a potential light verb,
rather than a “heavy” verb, if she recognizes that the verb can take an eventive noun as its direct
object argument. This proposal, however, seems to present a chicken-and-egg problem: On one
hand, children learn a noun is eventive (rather than object-denoting) from light verb constructions
(as discussed in Section II); but on the other hand, we propose that the recognition of light verb
constructions rely on event nominals. There might be a breach to this circle.
IV. The circle of event nominal and light verb acquisition: “do” breaks the path
This breach to this problem may be “do”––a light verb free of the ambiguity other light verbs have.
Although they used a combination of parental speech and the speech of the child her/himself; so, this is not precisely
a reflection of the input to the child, but nevertheless does capture the language in the child’s environment.
The verb do subcategorizes for an event (Bruening, 2019; Méry, 2002; Wamper, in prep), such
as (8a) and (8b), and hence is almost always used in light verb constructions (rarely in non-light
ones). Even when an object-denoting noun fills the direct argument of do, some eventive
meaning is usually coerced––for example, (8c) sounds very odd if used out of the blue, but may
be coerced to mean some activity associated with book”, as in (8d). Given this clean distribution
of VPs headed by “do”, the child learner does not have the problem of determining whether it is
the light or non-light. Therefore, “do” might be a path-breaking verb among all light verbs.
(8) a. Cal did a jump.
b. Cal did a rinse.
c. #Cal did a book.
d. To prepare for the presentation, we each produced some materials: Sue did a
newspaper; Joe did an NPR interview; and Cal did a book.
Once the child acquires the verb “do”, she may assume that the VP headed by “do” is a
light verb construction. From there, she infers that the direct object argument of that VP is an
event-denoting noun. Therefore, the child may use the “do”-constructions to learn a few event
nominals, and then use those few learned event nominals to learn other light verbs. From there,
she recognizes potential light verbs, and decide whether a potential light verb is used in its light
form or non-light form based on argument-role mappings. See Figure 3 for a summary of these
A caveat, however, is that eventive nouns can also be direct objects of non-light verbs, as in “imagine the kiss”, “like
the party”, etc. Analyses of the language input on the frequency of such occurrences are called for.
This proposal predicts that the child learner acquires the verb do before other light verbs.
Evidence from Barner's (2001) study on production shows that among the five light verbs
examined (have, give, take, do, and get), the use of do was most frequent at both an
earlier (MLU 1.0-2.0) and a later stage (MLU 3.0-4.0) of examination. Our corpus analysis results,
previously discussed in Section II (Figure 1), also show that in children’s production “do occur
most frequently as a light verb than other frequent verbs. What is interesting is that children use
“do” with some objects-denoting nouns (“we did the peas”, “do boots”), which rarely occur in
adult speech.
These appear to be coercions––that they simply use “do” as a verbalizer (Butt,
Future work should explore what drives them to use “do” in this way, despite lack of data
in the input. The finding from He & Arunachalam (under review) that children are able to use “do”
constructions (e.g., “I can do a gorp”) to infer that a novel noun is event- rather than object-
denoting, as discussed in Section II (p. 10), also suggests that “do” is understood early.
These data, however, only constitute indirect evidence. Future work on the acquisition
timeline as well as the input of “do” is called for. Empirical evidence directly testing the
prediction––that “do” is acquired earlier than other light verbs––needs to be gathered. Corpus
analyses of the use of “do” in child-directed speech (as well as in adult speech) should be
conducted to provide empirical support for the claim that “do” almost exclusively occur with nouns
that denote events. For both the above endeavors, attention should be paid to two points. First,
This finding is in line with a recent report that children (typically-developing as well as with SLI) have a tendency
to use general all-purpose verbs (“GAP” verbs) to substitute verbs with specific meanings (e.g., use “make a house”
instead of “build”, “make a hole” instead of “drill”) (Kambanaros & Grohmann, 2015).
Included in the count are also objects that seem to be used as locations or goals, without prepositions (e.g., “what
dat that do treasure chest”), creation usages (“you did the eyes and she did the nose”), and do as verb of imitation
(“watch me do horsie”).
auxiliary uses of “do” (e.g., “he didn’t jump”, “did she sleep?”, “Bill swam and so did I”) should
be distinguished from its light verb uses. In fact, the auxiliary uses of “do” may be another door
through which children come to recognize “do” is special; tracking and comparing auxiliary and
light-verb uses of “do”, therefore, will be useful. Second, idiomatic (or routine, non-analytic) uses
should be noted (e.g., “do me a favor”). Idiomatic phrases may be treated as a whole “chunk”
rather than learned analytically; but our point of learning event nominals from light verb
constructions require analytic encoding of the phrases (e.g., a light verb, the direct argument of
which is an event-denoting noun phrase). While idiomatic uses are common, their ratio relative to
non-idiomatic uses may be small, but empirical evidence would be preferred to back up this point.
Last but not least, another direction for future work that may shed light on our understanding of
the apparent circle and a potential “pathbreaker” lies in cross-linguistic investigation: if there is a
generic (i.e. unambiguously light) form of light verbs cross-linguistically, this may be part of the
child learner’s prior expectation, and the question boils down to finding the surface realization of
the expected category.
To conclude, in this paper, we have discussed the challenge of event nominals in the context
of syntax-semantics mappings––children assume nouns to name objects and use this assumption
as basis to infer novel noun meanings; how do they, then, learn nouns that denote event concepts?
Recent evidence suggests that light verb constructions might be an important cue. But light verb
constructions themselves are not straightforwardly easy to learn: they are late appearing in
production, and even their recognition presents challenges. The current literature does not yet have
clear answers; in fact, the challenges per se have not received enough attention.
We propose that the acquisition of event nominals and that of light verb constructions are
inter-related: Children acquire do first, as a member of the light verb category, and learn a few
event nominals from do-constructions; consequently, they may use those learned event nominals
to learn other light verbs in their language, and use the newly-learned light verbs as footsteps to
more event nominals. This proposal has an interesting implication to the big picture of syntax-
semantics mappings: event nominals are a case of mismatch between form and meaning at the
lexical level, light verb constructions are such a case at the clausal level, and they are cues to each
other. What is apparently a mismatch is in some sense a match.
Allerton, D. J. (2002). Stretched verb constructions in English. London, England: Routledge.
Ambridge, B., Pine, J. M., Rowland, C. F., Freudenthal, D., & Chang, F. (2014). Avoiding dative
overgeneralisation errors: Semantics, statistics or both? Language, Cognition and
Neuroscience, 29(2), 218243.
Arunachalam, S., & He, A. X. (2018). Children’s acquisition of nouns that denote events. In
Proceedings of the 42nd Boston University Conference on Language Development
(BUCLD) (pp. 2944). Somervillle, MA: Cascadilla Press.
Arunachalam, S., & Waxman, S. R. (2010). Meaning from syntax: Evidence from 2-year-olds.
Cognition, 114(3), 442446.
Baker, M. C. (1989). Object sharing and projection in serial verb constructions. Linguistic
Inquiry, 20, 513553.
Barner, D. (2001). Light verbs and the flexible use of words in early language learning. Master’s
thesis, McGill University.
Bernal, S., DehaeneLambertz, G., Millotte, S., & Christophe, A. (2010). Twoyearolds
compute syntactic structure online. Developmental Science, 13(1), 6976.
Bernal, S., Lidz, J., Millotte, S., & Christophe, A. (2007). Syntax constrains the acquisition of
verb meaning. Language Learning and Development, 3(4), 325341.
Bloom, P. (1990). Subjectless sentences in child language. Linguistic Inquiry, 21, 491504.
Borer, H. (1999). Deconstructing the construct. In K. Johnson & Roberts I. (Eds.), Beyond
principles and parameters (pp. 4389). Dordrecht: Kluwer.
Borer, H., & Wexler, K. (1987). The maturation of syntax. In T. Roeper & E. Williams (Eds.),
Parameter setting. Studies in theoretical psycholinguistics, vol 4 (pp. 123172). Dordrecht:
Brown, P. (2008). Verb specificity and argument realization in Tzeltal child language. In M.
Bowerman & P. Brown (Eds.), Crosslinguistic perspectives on argument structure:
Implications for learnability (pp. 167190). New York: Oxford University Press.
Brown, R. W. (1957). Linguistic determinism and the part of speech. The Journal of Abnormal
and Social Psychology, 55(1), 15.
Bruening, B. (2019). Passive do so. Natural Language and Linguistic Theory, 37(1), 149.
Brugman, C. (2001). Light verbs and polysemy. Language Sciences, 23, 551578.
Butt, M. (2003). The light verb jungle. Harvard Working Papers in Linguistics, 9, 149.
Butt, M. (2010). The light verb jungle: Still hacking away. In M. Amberber, M. Harvey, & B.
Baker (Eds.), Complex predicates in cross-linguistic perspective (pp. 4878). Cambridge:
Cambridge University Press.
Butt, M., & Lahiri, A. (2013). Diachronic pertinacity of light verbs. Lingua, 125, 729.
Casati, R., & Varzi, A. C. (2008). Event concepts. In T. F. Shipley & J. Zacks (Eds.),
Understanding events: From perception to action (pp. 3154). New York: Oxford
University Press.
Cattell, R. (1984). Composite predicates in English. New York: Academic Press.
Chomsky, N. (1981). Lectures on government and binding. Dordrecht: Foris.
Clark, E. V. (1978). From gesture to word: On the natural history of deixis in language
acquisition. In J. S. Bruner & A. Garton (Eds.), Human growth and development (pp. 85
120). London, England: Oxford University Press.
de Carvalho, A., Dautriche, I., Lin, I., & Christophe, A. (2017). Phrasal prosody constrains
syntactic analysis in toddlers. Cognition, 163, 6779.
De Villiers, J. (1985). Learning how to use verbs: Lexical coding and the influence of the input.
Journal of Child Language, 12(12), 587595.
Dowty, D. R. (1989). On the semantic content of the notion of ‘thematic role.’ In Properties,
types and meaning (pp. 69129). Dordrecht: Springer.
Dowty, D. R. (1991). Thematic proto-roles and argument selection. Language, 67, 547619.
Echols, C. H., & Marti, C. N. (2004). The identification of words and their meanings: From
perceptual biases to language-specific cues. In D. G. Hall & S. R. Waxman (Eds.), Weaving
a lexicon (pp. 4178). Cambridge, MA: MIT Press.
Everaert, M., & Hollebrandse, B. (1995). The lexical representation of light verb constructions.
In Proceedings of the Twenty-First Annual Meeting of the Berkeley Linguistics Society:
General Session and Parasession on Historical Issues in Sociolinguistics/Social Issues in
Historical Linguistics (pp. 94104). Berkeley: Berkeley Linguistics Society.
Fennell, C. T. (2006). Infants of 14 months use phonetic detail in novel words embedded in
naming phrases. In D. Banman, T. Magnitsikai, & C. Zaller (Eds.), Proceedings of the 30th
annual Boston University Conference on Language Development (pp. 178189).
Somerville, MA: Cascadilla Press.
Fenson, L., Dale, P. S., Reznick, J. S., Thal, D. J., Bates, E., Hartung, J. P., … Reilly, J. S.
(1993). The MacArthur Communicative Development Inventories: User’s guide and
technical manual. Baltimore, MD: Paul H. Brookes Pub. Co.
Fisher, C. (1996). Structural limits on verb mapping: The role of analogy in children’s
interpretations of sentences. Cognitive Psychology, 31(1), 4181.
Fisher, C., Hall, D. G., Rakowitz, S., & Gleitman, L. R. (1994). When it is better to receive than
to give: Syntactic and conceptual constraints on vocabulary growth. Lingua, 92, 333375.
Gerken, L. (1991). The metrical basis for children’s subjectless sentences. Journal of Memory
and Language, 30(4), 431451.
Gleitman, L. R. (1990). The structural sources of verb meanings. Language Acquisition, 1(1), 3
Göksun, T., Küntay, A. C., & Naigles, L. R. (2008). Turkish children use morphosyntactic
bootstrapping in interpreting verb meaning. Journal of Child Language, 35, 291323.
Goldberg, A. E. (1998). Patterns of experience in patterns of language. In M. Tomasello (Ed.),
The new psychology of language (pp. 203218). Mahwah, NJ: Lawrence Erlbaum.
Goldberg, A. E. (2013). The emergence of the semantics of argument structure constructions. In
The Emergence of Language (pp. 215230). New York: Psychology Press.
Grimshaw, J. (1981). Form, function and the language acquisition device. In C. L. Baker & J. J.
McCarthy (Eds.), The logical problem of language acquisition. Cambridge, MA: MIT
Grimshaw, J. (1990). Argument structure. Cambridge, MA: MIT Press.
He, A. X., & Arunachalam, S. (under review). When nouns do not name objects: Todders’
acquisition of event nominal words.
He, A. X., & Lidz, J. (2017). Verb learning in 14- and 18-month-old English-learning infants.
Language Learning and Development, 13(3), 335356.
He, A. X., Wellwood, A., Lidz, J., & Williams, A. (2013). Assessing event perception in adults
and prelinguistic children: A prelude to syntactic bootstrapping. In Poster presented at the
38th Boston University Conference on Language Development (BUCLD). Boston
University, Boston, MA.
Hicks, J. (2006). The impact of function words on the processing and acquisition of syntax.
Linguistics. Doctoral dissertation, Northwestern University.
Hirsch, C. K. (2011). The acquisition of raising. Doctoral dissertation, Massachusetts Institute of
Hirsh-Pasek, K., Golinkoff, R. M., Fletcher, P., DeGaspe-Beaubien, F., & Cauley, K. M. (1985).
In the beginning: One-word speakers comprehend word order. In Paper presented at Boston
University Conference on Language Development (BUCLD). Boston, MA.
Höhle, B., Weissenborn, J., Kiefer, D., Schulz, A., & Schmitz, M. (2004). Functional elements in
infants’ speech processing: The role of determiners in the syntactic categorization of lexical
elements. Infancy, 5(3), 341353.
Jackendoff, R. (1974). A deep structure projection rule. Linguistic Inquiry, 5(4), 481505.
Jackendoff, R. (2002). The complexity of linguistic structure. In Foundations of language:
Brain, meaning, grammar, evolution. New York: Oxford University Press.
Jespersen, O. (1954). A modern English grammar on historical principles. London, England:
Allen & Unwin and Copenhagen: Ejnar Munksgaard.
Jusczyk, P. W., & Aslin, R. N. (1995). Infants′ detection of the sound patterns of words in fluent
speech. Cognitive Psychology, 29(1), 123.
Kambanaros, M., & Grohmann, K. K. (2015). More general all-purpose verbs in children with
specific language impairment? Evidence from Greek for not fully lexical verbs in language
development. Applied Psycholinguistics, 36, 10251057.
Keenan, E. L. (1976). Towards a universal definition of “subject.” In C. Li (Ed.), Subject and
topic. New York: Academic Press.
Kirby, S. (2009). Semantic scaffolding in first language acquisition: The acquisition of raising-
to-object and object control. Doctoral dissertation, University of North Carolina at Chapel
Kobayashi, T., Mugitani, R., & Amano, S. (2006). Development of infants ability to associate
words with intrinsic motions. Poster Presented at XVth Biennial International Conference
on Infant Studies. Westin Miyako, Kyoto, Japan.
Landau, B., & Gleitman, L. R. (1985). Language and experience: Evidence from the blind child.
Cambridge, MA: Harvard University Press.
Lee, J. N., & Naigles, L. R. (2008). Mandarin learners use syntactic bootstrapping in verb
acquisition. Cognition, 106(2), 10281037.
Lidz, J., Gleitman, H., & Gleitman, L. R. (2003). Understanding how input matters: Verb
learning and the footprint of universal grammar. Cognition, 87(3), 151178.
Lieven, E., Behrens, H., Speares, J., & Tomasello, M. (2003). Early syntactic creativity: A
usage-based approach. Journal of Child Language, 30(2), 333370.
Macnamara, J. T. (1982). Names for things: A study of human learning. Cambridge, MA: MIT
MacWhinney, B. (2000). The CHILDES Project: Tools for Analyzing Talk (third edition)
Volume I: Transcription format and programs Volume II: The database. Mahwah, NJ:
Lawrence Erlbaum Associates.
Maouene, J., Laakso, A., & Smith, L. B. (2011). Object associations of early-learned light and
heavy English verbs. First Language, 31(1), 109132.
Matsuo, A., Kita, S., Shinya, Y., Wood, G. C., & Naigles, L. (2012). Japanese two-year-olds use
morphosyntax to learn novel verb meanings. Journal of Child Language, 39(3), 637.
Méry, R. (2002). A propos du verbe DO. Anglophonia, 12, 81133.
Mintz, T. H. (2003). Frequent frames as a cue for grammatical categories in child directed
speech. Cognition, 90(1), 91117.
Mintz, T. H. (2006). Finding the verbs: Distributional cues to categories available to young
learners. In K. Hirsh-Pasek & R. M. Golinkoff (Eds.), Action meets word: How children
learn verbs (pp. 3163). New York: Oxford University Press.
Moulton, K. (2014). Simple event nominalizations: Roots and their interpretation. In I. Paul
(Ed.), Cross-linguistic investigations of nominalizations patterns (pp. 119144).
Amsterdam: John Benjamins.
Müller, S. (2016). Grammatical theory: From transformational grammar to constraint-based
approaches. Berlin, Germany: Language Science Press.
Naigles, L. R. (1990). Children use syntax to learn verb meanings. Journal of Child Language,
17(02), 357374.
Naigles, L. R., & Hoff-Ginsberg, E. (1998). Why are some verbs learned before other verbs?
Effects of input frequency and structure on children’s early verb use. Journal of Child
Language, 25(1), 95120.
Naigles, L. R., & Kako, E. T. (1993). First contact in verb acquisition: Defining a role for syntax.
Child Development, 64(6), 16651687.
Nelson, K., Hampson, J., & Shaw, L. K. (1993). Nouns in early lexicons: Evidence, explanations
and implications. Journal of Child Language, 20(1), 6184.
Ninio, A. (1999). Pathbreaking verbs in syntactic development and the question of prototypical
transitivity. Journal of Child Language, 26(3), 619653.
Oshima-Takane, Y., Ariyama, J., Kobayashi, T., Katerelos, M., & Poulin-Dubois, D. (2011).
Early verb learning in 20-month-old Japanese-speaking children. Journal of Child
Language, 38(3), 455484.
Oshima-Takane, Y., Barner, D., Elsabbagh, M., & Guerriero, S. (2001). Leaming of deverbal
nouns. In M. Almgren, A. Barreña, M.-J. Ezeizabarrena, I. Idiazabal, & B. MacWhinney
(Eds.), Research on Child Language Acquisition: Proceedings for the 8th Congress of the
International Association for the Study of Child Language. Somerville, MA: Cascadilla
Press. Retrieved from
Oshima-Takane, Y., Satin, J., & Tint, A. (2008). Rapid word-action mapping in French- and
English-speaking children. In H. Chan, H. Jacob, & E. Kapia (Eds.), 32nd Annual Boston
University Conference on Language Development (pp. 347359). Somerville, MA:
Cascadilla Press.
Piñango, M. M., Mack, J., & Jackendoff, R. (2006). Semantic combinatorial processes in
argument structure: Evidence from light-verbs. Annual Meeting of the Berkeley Linguistics
Society, 32(1), 573583.
Pinker, S. (1984). The semantic bootstrapping hypothesis. In Language learnability & language
development (pp. 303333). Cambridge, MA: Harvard University Press.
Pinker, S. (1989). Learnability and cognition: The acquisition of argument structure.
Cambridge: MIT Press.
Prudent, S. M., Hirsh-Pasek, K., Golinkoff, R. M., & Hennon, E. A. (2006). The birth of words:
Tenmontholds learn words through perceptual salience. Child Development, 77(2), 266
Quine, W. van O. (1960). Word and object. Cambridge, MA: MIT press.
Ramchand, G. C. (2013). On structural meaning vs. conceptual meaning in verb semantics.
Linguist Analysis, 39(1), 207244.
Resnik, P. (1996). Selectional constraints: An information-theoretic model and its computational
realization. Cognition, 61(12), 127159.
Rissman, L., Rawlins, K., & Landau, B. (2015). Using instruments to understand argument
structure: Evidence for gradient representation. Cognition, 142, 266290.
Roy, I., & Soare, E. (2013). Event-related nominals. In Categorization and category change (pp.
123152). Newcastle upon Tyne, England: Cambridge Scholars Publishing.
Santelmann, L. M., & Jusczyk, P. W. (1998). Sensitivity to discontinuous dependencies in
language learners: Evidence for limitations in processing space. Cognition, 69(2), 105134.
Shi, R. (2014). Functional morphemes and early language acquisition. Child Development
Perspectives, 8(1), 611.
Theakston, A., Lieven, E., Pine, J., & Rowland, C. (2004). Semantic generality, input frequency
and the acquisition of syntax. Journal of Child Language, 31(1), 6199.
Tomasello, Michael. (2000). The item-based nature of children’s early syntactic development.
Trends in Cognitive Sciences, 4(4), 156163.
Tomasello, Michael. (2003). Constructing a language: A usage-based theory of language
acquisition. Cambridge, MA: Harvard University Press.
Trehub, S. E., & Shenfield, T. (2007). Acquisition of early words from single-word and
sentential contexts. Developmental Science, 10(2), 190198.
Valian, V. (1991). Syntactic subjects in the early speech of American and Italian children.
Cognition, 40(12), 2181.
Wamper, J. (n.d.). We also do dogs: Event representation from context.
Waxman, S. R., & Booth, A. E. (2001). Seeing pink elephants: Fourteen-month-olds’
interpretations of novel nouns and adjectives. Cognitive Psychology, 43(3), 217242.
Waxman, S. R., Lidz, J., Braun, I. E., & Lavin, T. (2009). Twenty four-month-old infants’
interpretations of novel verbs and nouns in dynamic scenes. Cognitive Psychology, 59(1),
Wellwood, A., He, A. X., Lidz, J., & Williams, A. (2015). Participant structure in event
perception: Towards the acquisition of implicitly 3-place predicates. In University of
Pennsylvania Working Papers in Linguistics (Vol. 21, p. Article 32). Philadelphia, PA:
University of Pennsylvania. Retrieved from
Wexler, K. (1998). Very early parameter setting and the unique checking constraint: A new
explanation of the optional infinitive stage. Lingua, 106(14), 2379.
Wiese, H. (2006). “Ich mach dich Messer”: Grammatische Produktivität in Kiez-Sprache.
Linguistische Berichte, 207, 245273.
Williams, A. (2008). Patients in Igbo and Mandarin. In J. Dlling, T. Heyde-Zybatow, & M.
Schäfer (Eds.), Event structures in linguistic form and interpretation (pp. 330). Berlin:
Walter de Gruyter.
Williams, A. (2015). Arguments in syntax and semantics. Cambridge: Cambridge University
Wittenberg, E., Khan, M., & Snedeker, J. (2017). Investigating thematic roles through implicit
learning: Evidence from light verb constructions. Frontiers in Psychology, 8.
Wittenberg, E., & Levy, R. (2017). If you want a quick kiss, make it count: How choice of
syntactic construction affects event construal. Journal of Memory and Language, 94, 254
Wittenberg, E., Paczynski, M., Wiese, H., & Kuperberg, G. R. (2014). The difference between
“giving a rose” and “giving a kiss”: Sustained neural activity to the light verb construction.
Journal of Memory and Language, 73, 3142.
Wittenberg, E., & Piñango, M. M. (2011). Processing light verb constructions. The Mental
Lexicon, 6, 393413.
Wittenberg, E., & Snedeker, J. (2014). It takes two to kiss, but does it take three to give a kiss?
Categorization based on thematic roles. Language, Cognition and Neuroscience, 29(5),
Yuan, S., & Fisher, C. (2009). “Really? She blicked the baby?” Two-year-olds learn
combinatorial facts about verbs by listening. Psychological Science, 20(5), 619626.
Yuan, S., Fisher, C., & Snedeker, J. (2012). Counting the nouns: Simple structural cues to verb
meaning. Child Development, 83(4), 13821399.
Figure Legends
Figure 1. Eventive/light (green) and non-eventive/non-light (blue) syntactic objects used with
verbs that can function as light verbs in the Brown corpus, broken down by verb (LEFT) and age
of children (RIGHT).
Figure 2. Syntax-semantics mappings for non-light (a) and light (b) constructions
Figure 3. Inferences between event nominals and light verbs, with “do” to break the circle (Ni
and Nk are specific event nominals, and Va a specific verb)
... For such an utterance, syntax is irrelevant -there is no point to assigning syntactic categories like noun or verb, or their syntactic markers such as determiners or verbal inflection. The utterances are interpreted based on their meaning alone -if a child says 'nap', she is talking about an event; it is irrelevant whether 'nap' in this case is a verb or noun (Arunachalam & He, 2018;He & Wittenberg, 2020). These interpretations would not be possible in a grammar like (4). ...
Full-text available
We propose a Complexity Hierarchy of grammars that map between sound and meaning, beginning with relatively trivial one-word grammars and culminating with the grammars of modern human languages. We argue that the levels in this hierarchy are plausible and necessary stages in the evolution of the contemporary human language faculty. We further argue that these levels leave traces in syntactically complex languages, as well as in home sign and emerging sign languages, where they can be thought of as "living fossils." An important stage in the hierarchy is what has been called "protolanguage" or "linear grammar"-a form of language that maps meaning to strings of words, but that, crucially, lacks syntactic and morphological structure. This gradualist scenario has implications for the division of labor between grammar and pragmatics. The simpler grammars in the Complexity Hierarchy place a strong reliance on pragmatics for many aspects of utterance meaning, including even basic questions such as who did what to whom. As grammars move up the Complexity Hierarchy and become more complex, these relatively simple interpretive factors become more systematic and less dependent on pragmatic inferences. However, pragmatic processes do not disappear. Rather, they change in character: Syntax, semantics, and the lexicon trigger highly structured pragmatic phenomena such as presuppositions and implicatures in a systematic and reliable way. In a sense, the more complex the grammar, the more opportunity for such pragmatic niches.
... For such an utterance, syntax is irrelevant -there is no point to assigning syntactic categories like noun or verb, or their syntactic markers such as determiners or verbal inflection. The utterances are interpreted based on their meaning alone -if a child says 'nap', she is talking about an event; it is irrelevant whether 'nap' in this case is a verb or noun (Arunachalam & He, 2018;He & Wittenberg, 2020). These interpretations would not be possible in a grammar like (4). ...
We propose a Complexity Hierarchy of grammars that map between sound and meaning, beginning with relatively trivial one-word grammars and culminating with the grammars of modern human languages. We argue that the levels in this hierarchy are plausible and necessary stages in the evolution of the contemporary human language faculty. We further argue that these levels leave traces in syntactically complex languages, as well as in home sign and emerging sign languages, where they can be thought of as "living fossils." An important stage in the hierarchy is what has been called "protolanguage" or "linear grammar"-a form of language that maps meaning to strings of words, but that, crucially, lacks syntactic and morphological structure. This gradualist scenario has implications for the division of labor between grammar and pragmatics. The simpler grammars in the Complexity Hierarchy place a strong reliance on pragmatics for many aspects of utterance meaning, including even basic questions such as who did what to whom. As grammars move up the Complexity Hierarchy and become more complex, these relatively simple interpretive factors become more systematic and less dependent on pragmatic inferences. However, pragmatic processes do not disappear. Rather, they change in character: Syntax, semantics, and the lexicon trigger highly structured pragmatic phenomena such as presuppositions and implicatures in a systematic and reliable way. In a sense, the more complex the grammar, the more opportunity for such pragmatic niches.
Full-text available
In this paper we present a study about the typical development of the comprehension of expressions that exhibit an ambiguity between a literal and a nonliteral interpretation in Spanish, and whose most frequent use is nonliteral. Such expressions include light verb constructions (LVC) such as to make the bed and expressions in a metaphor‐hyperbole‐idiom continuum (MHI) such as to sleep with angels . We ran a forced‐choice experiment where children aged 3 to 9 (N = 143) heard an ambiguous expression and had to choose the correct picture on the face of three options: one target item and two distractors. There were two counterbalanced lists, so that each critical item would be present in either the literal or the nonliteral condition. We collected accuracy data as well as reaction times. We encountered different developmental trends for LVC than for MHI, observing a literalist stage in MHI which we did not observe in LVC.
Full-text available
We experimentally investigate the meaning of denominal verbs in child and adult Romanian using a semi-artificial/semi-nonce denominal verb (SAD) paradigm, i.e., using non-existent verbs derived from familiar nouns (a cireşi ‘to cherry’). Importantly, the SAD paradigm allows us to probe into meaning formation without the lexical bias of existing verbs. To see whether children have difficulties understanding SAD verbs in linguistic contexts, we conducted a Contextual Denominal Task. Children were asked to select a matching picture after hearing sentences with SAD verbs in linguistic contexts biasing them for a particular interpretation. Children generally opted for a literal interpretation of a cireşi ‘to cherry’, involving the actual object cherry (‘to pick/eat cherries’), over a figurative interpretation such as a deveni (roşie) ca cireaşa ‘to become (red) like a cherry’, i.e., ‘to blush’ even in figurative-biasing contexts (like Mary cherried when John told her she was beautiful). In order to see whether children perform better when the meaning is made explicit or whether they have a general difficulty with figurative meanings (whether implicit or explicit), we also conducted an Explicit Denominal Paraphrase Task, where children were instead exposed to the corresponding denominal paraphrases (e.g., a deveni ca cireaşa ‘to become like a cherry’). Children performed almost adult-like when the figurative meaning was more explicit. We account for our findings within a Meaning First Approach (Sauerland & Alexiadou 2020; Guasti, Alexiadou & Sauerland 2023), which assumes that compressed meaning is hard, and that decompressing words is subject to two possible principles: (structural and conceptual) simplicity and plausibility. While adults tend to observe plausibility, children prefer simplicity more, generally opting for literal readings, which merge the light verb DO or similar verbs with nouns (Hale & Keyser 2002; Kiparsky 1997).
Full-text available
The present paper introduces a radical embodied analysis of German modals. It is argued that German modals reflect a connection between perception, memory and consciousness as a means to reconstruct reality. This connection, termed herein "extended mind" aims at providing a renewed characterization of language as a semiotic tool that bears traces of somatosensory information for the reconstruction specific events. In particular, it is claimed that modal events in German (and other languages) are constructed through recourse to embodied information encoded by specific constructional attachment patterns. Thus, instead of constructions, the claim is that humans make use of stored perceptual maps, the combination of which allows us to manipulate complex event partitioning.
Full-text available
Light verb constructions (e.g. give a sigh, take a walk) are a linguistic puzzle, as they consist of two predicating elements in a monoclausal structure. In the theoretical literature, there has been much interest in the linguistic analysis of such constructions across a range of grammatical frameworks. One such proposal is event co-composition, where the argument structures of noun and light verb merge, resulting in a composite argument structure, which has been claimed to be the source of increased processing costs in English and German. In contrast to these languages, in Hindi a larger proportion of the predicates are light verb constructions. Hence, we may ask whether a Hindi speaker's experience with light verb constructions allows them to go through the same co-composition operation faster than a speaker of English. Our results show that Hindi speakers are adept at the process of using light verb constructions to 'verbalize' predicates, more so than speakers of Germanic languages. We argue that these data provide evidence for a case of specific linguistic experiences shaping cognition: cost disappears with practice.
Full-text available
The received view is that the VP pro-form do so cannot be a verbal passive, although it can be unaccusative. I show that this is incorrect: do so can be passive. It can also take a raising to subject verb as its antecedent. This means that do so is compatible with all types of A-movement, although it does not permit A-bar movement of an object. I construct an analysis of do so where it is simply an intransitive verb plus an adverb. Do so combines with a Voice head, which can be unaccusative, passive, or active transitive. The subject of do so is base-generated in Spec-VoiceP, and does not move in unaccusatives or passives. Instead, do so must copy a function from its antecedent in the semantics. The subject of do so can be interpreted semantically as an internal argument if Voice is passive or unaccusative and the antecedent includes a trace, because of the way lambda abstraction works in A-movement. This analysis reconciles the evidence against movement in do so itself with arguments for A-movement in its antecedent. The copy mechanism explains voice and category mismatches, as well as split antecedents and ellipsis-containing antecedents.
Full-text available
The syntactic structure of a sentence is usually a strong predictor of its meaning: Each argument noun phrase (i.e., Subject and Object) should map onto exactly one thematic role (i.e., Agent and Patient, respectively). Some constructions, however, are exceptions to this pattern. This paper investigates how the syntactic structure of an utterance contributes to its construal, using ditransitive English light verb constructions, such as “Nils gave a hug to his brother”, as an example of such mismatches: Hugging is a two-role event, but the ditransitive syntactic structure suggests a three-role event. Data from an eye-tracking experiment and behavioral categorization data reveal that listeners learn to categorize sentences according to the number of thematic roles they convey, independent of their syntax. Light verb constructions, however, seem to form a category of their own, in which the syntactic structure leads listeners down an initial incorrect assignment of thematic roles, from which they only partly recover. These results suggest an automatic influence of syntactic argument structure on semantic interpretation and event construal, even in highly frequent constructions.
Full-text available
The present study investigates English-learning infants’ early understanding of the link between the grammatical category verb and the conceptual category event, and their ability to recruit morphosyntactic information online to learn novel verb meanings. We report two experiments using an infant-controlled Habituation-Switch Paradigm. In Experiment 1, we habituated 14- and 18- month-old infants with two scenes each labeled by a novel intransitive verb embedded in the frame “is ___ing”: a penguin-spinning scene paired with “it’s doking”, a penguin-cartwheeling scene paired with “it’s pratching”. At test, infants in both age groups dishabituated when the scene-sentence pairings got switched (e.g., penguin-spinning—“it’s pratching”). This finding is consistent with two explanations: (1) infants were able to link verbs to event concepts (as opposed to other concepts, e.g., objects) and (2) infants were simply tracking the surface-level mapping between scenes and sentences, and it was scene-sen- tence mismatch that elicited dishabituation, not their knowledge of verb-event link. In Experiment 2, we investigated these two possibilities, and found that 14- month-olds were sensitive to any type of mismatch, whereas 18-month-olds dishabituated only to a mismatch that involved a change in word meaning. Together, these results provide evidence that 18-month-old English-learning infants are able to learn novel verbs by recruiting morphosyntactic cues for verb categorization and use the verb-event link to constrain their search space of possible verb meanings. [To access the final version (limited copies), please use this link:]
This study examined whether phrasal prosody can impact toddlers’ syntactic analysis. French noun-verb homophones were used to create locally ambiguous test sentences (e.g., using the homophone as a noun: [le bébé souris] [a bien mangé] - [the baby mouse] [ate well] or using it as a verb: [le bébé] [sourit à sa maman] - [the baby] [smiles to his mother], where brackets indicate prosodic phrase boundaries). Although both sentences start with the same words (le-bebe-/suʁi/), they can be disambiguated by the prosodic boundary that either directly precedes the critical word /suʁi/ when it is a verb, or directly follows it when it is a noun. Across two experiments using an intermodal preferential looking procedure, 28-month-olds (Exp. 1 and 2) and 20-month-olds (Exp. 2) listened to the beginnings of these test sentences while watching two images displayed side-by-side on a TV-screen: one associated with the noun interpretation of the ambiguous word (e.g., a mouse) and the other with the verb interpretation (e.g., a baby smiling). The results show that upon hearing the first words of these sentences, toddlers were able to correctly exploit prosodic information to access the syntactic structure of sentences, which in turn helped them to determine the syntactic category of the ambiguous word and to correctly identify its intended meaning: participants switched their eye-gaze toward the correct image based on the prosodic condition in which they heard the ambiguous target word. This provides evidence that during the first steps of language acquisition, toddlers are already able to exploit the prosodic structure of sentences to recover their syntactic structure and predict the syntactic category of upcoming words, an ability which would be extremely useful to discover the meaning of novel words.