Cognitive Mechanisms Underlying Fake Reflexive Resultatives^*

* We would like to thank two anonymous reviewers for their feedback. This work was supported by the Spanish Ministry of Economy and Competitiveness [Grants numbers FFI2011-29798-C02-01 and FFI2013-43593-P].

Abstract

Both projectionist and constructional approaches to language support competing arguments as regards the place where grammar meets the lexicon. While the former claim that the morphosyntactic realization of verbal arguments is determined by the lexical semantic representation of the verb, the latter hold that lexicon and grammar form a continuum and put forward a shift of emphasis from the pivotal role of the verb within the sentence to the notion of construction as a form/meaning or function pairing. Moreover, constructionists stipulate that lexical–constructional fusion is regulated by some constraints (e.g. Goldberg’s semantic constraints or Michaelis’ Override Principle). We provide further evidence in favour of the constructional approach through the analysis of the fake reflexive resultative construction. We concur with Goldberg and Levin that semantically similar verbs tend to participate in the same argument structure constructions. To systematize our analysis, we take as a basis Levin’s classification of verbs and their distribution across Halliday and Matthiessen’s process types. Additionally, we make a contribution to the literature on constraints on lexical–constructional fusion by discussing some cognitive mechanisms, mainly high-level metaphor and metonymy, which license or block out this process.

Keywords:

• Fake Reflexives
• High-level Metaphor
• High-level Metonymy
• Transitivity System
• Motivation

Notes

* We would like to thank two anonymous reviewers for their feedback. This work was supported by the Spanish Ministry of Economy and Competitiveness [Grants numbers FFI2011-29798-C02-01 and FFI2013-43593-P].

1 Levin (1993) deals with different groups of lexical verbs and briefly discusses the constructions which prototypically are licensed by each set of predicates. However, she does not provide a very exhaustive analysis of each set of verbs and the motivational mechanisms which underlie the possibility of lexical–constructional fusion. This is mainly due to Levin's focus on the whole English lexicon and it is undoubtedly a very exhaustive and excellent starting point for the kind of research presented in this proposal.

2 Cognitive linguists also argue for the fuzzy nature of categories. The higher the degree of resemblance of a member to the prototype, the better an example of the category to which the member belongs it will be (see Aarts et al. (2004) for detailed information on prototypicality and fuzzy categories). The system of transitivity (and the linguistic system as a whole) abides by the Principle of Systemic Indeterminacy, which stipulates that since the world of our experience is highly indeterminate, the indeterminacy in the system of process types is to be expected and naturally stems from our experience of the world around us (Halliday and Matthiessen 2004: 173).

3 Verbs like smile and nod can be used transitively. However, the object in occurrences like Sarah smiled a charming smile is a contribution of the cognate object construction, of the reaction object construction in examples such as He smiled his congratulations or He nodded his agreement, and of the understood body-part object construction in instances like He nodded his head. As has been observed (Eggins 2004: 233–234), the behaviour and the phenomenon are optional participants in behaviourals. Our examples include intransitive predicates or verbs which can be both transitive and intransitive, like smile and nod, but are used in their intransitive versions.

4 We are exclusively concerned with property fake reflexive resultatives (both AP and PP property realizations) in this section because the occurrence of PP spatial fake reflexive resultatives is almost negligible. For instance, by running the query laughed herself in the American version of Google Books (http://googlebooks.byu.edu/x.asp), we get 51 instances of the fake reflexive construction in the period from 2000 to 2009. Only three of them are spatial fake reflexive resultatives (e.g. She laughed herself onto the floor; Google Books: Dangerously Close, by Dee J. Adams, 2012. Accessed 10 November 2015). The query of the string sobbed himself in the same corpus in the same period yields 44 occurrences of fake reflexive resultatives, none of which are examples of spatial ones. And by looking up the search string sobbed herself, we gather no spatial fake reflexive resultatives from the set of 45 occurrences found either. In contrast to prepositional examples like (2), spatial realizations do not abide by the general metaphor STATES ARE LOCATIONS and by the more specific metaphor A CHANGE OF STATE IS A CHANGE OF LOCATION. For instance, in She laughed herself onto the floor, ‘she’ is the Behaver, seen as the Actor of a material contact process, who affects herself in such a way that the outcome is a change of location (from somewhere which was not onto the floor to somewhere onto the floor) and not a change of state. Additionally, by virtue of the high-level metaphor A MATERIAL CONTACT PROCESS IS CAUSED MOTION, the Actor is seen as the causer of motion, the predicate is mapped onto causing motion, the Goal is regarded as the object of motion and the Resultant Attribute is the location reached. And finally, the metonymy MANNER FOR BEHAVIOURAL PROCESS applies to this example as well, since manner affords conceptual access to the whole matrix domain of actions in such a way that this occurrence can be paraphrased as ‘She caused herself to move onto the floor by laughing’.

5 The importance allotted to pragmatic factors in constructions in general and in the resultative construction in particular has also been claimed by other scholars like Goldberg and Jackendoff (2004) and Pérez and Peña (2009).

6 Rappaport and Levin (2001: 779–782) discuss the optionality of the reflexive pronoun in the fake reflexive construction in syntactic terms. According to them, in this pattern, the presence of the reflexive pronoun coreferent with the subject is accounted for by the argument-per-subevent condition. This condition stipulates that for each subevent in the event structure there should be at least one argument XP in the syntax (Rappaport & Levin 2001: 779). The fake reflexive construction conveys complex events consisting of two subevents which have one participant in common, the item which is the subject of the resultative pattern. Thus, this shared participant is linguistically realized twice. In application of the argument-per-subevent condition, an object (the reflexive pronoun) is required. Moreover, Rappaport and Levin (2001) set the bare XP pattern (e.g. Robin danced out of the room) and the fake reflexive configuration (e.g. Robin danced herself out of the room) in contrast. These scholars argue that a bare resultative like Robin danced out of the room is preferred to a reflexive resultative such as Robin danced herself out of the room when the subevents satisfy the conditions for event identity. They add that this preference can be overridden by pragmatic factors.

7 This characteristic of fake reflexives is also predicated of the way-construction. Scholars like Christie (2011) and Mondorf (2011) have devoted their research to the similarities and differences between these two constructions since they seem to be feasible in almost the same contexts.

8 There is considerable detailed interesting discussion of the varied distribution of various to-phrases in a similar semantic domain in Grône (2014).

9 In this connection, Díez (2002: 311) put forward that actions can be regarded as domains of experience. Moreover, she stated that the relationship between processes and actions is a metonymic one since processes are actions which lack the role (or subdomain) of agent, which is characteristic of actions. We concur with her that any process can be understood as a domain which comprises subdomains like participants and circumstances.

10 Even though our examples (23)–(33) are transitive realizations, the syntactic objects are a contribution of the fake reflexive resultative construction, as was the case with the examples in Section 5.1.

11 Since chant is related in meaning to sing, we discuss them as belonging to the same group.

12 As was the case in the previous section and will be the case in Section 5.4, this section is only concerned with property fake reflexive resultatives since our analysis of the data reveals that there are very few occurrences of spatial fake reflexive resultatives which contain verbs involving the body, verbs of communication and verbs of ingesting.

13 Halliday and Matthiessen (2004: 187–189) include in their work a table with examples of verbs serving as process in different material clause types. Several subtypes are identified and each of them comprises a list of verbs and they are classified as intransitive or transitive. On closer inspection, we notice that contact verbs are always transitive while motion verbs can be both transitive and intransitive even though they show a clear tendency to be intransitive.

14 In contrast to the examples in Sections 5.1, 5.2 and 5.4, PP spatial fake reflexive resultatives are frequent. For instance, we have gathered 189 realizations of the fake reflexive construction by searching for the string ‘[worm] *self’ in the American version of Google Books. Of these, 123 examples are property resultatives (especially PP property instantiations) and 66 comply with the characteristics of PP spatial resultatives. This means a higher proportion of PP spatial fake reflexive resultatives than in the rest of the groups of verbs and results from the nature of the verbs in this section.

15 Of course, we mean metaphorical motion.

16 In fact, again the intransitive versions or uses of these verbs are considered and call for a process of transitivization in order to adapt to the requirements of material contact processes. Even though it might seem contradictory to pick out the intransitive version of a verb which can be transitive as well and then make it transitive through metaphorical activity, it is not. The nature of the possible Goals of both ingestion and material contact processes is very different. In material contact processes the Goal is physically affected by the action of the predicate, which is not the case with ingestion verbs, among others. Moreover, as was the case with verbs in Section 5.1, these expressions do not focus on what is eaten/drunk/chewed/gnawed/gulped but on the process itself and thanks to the metaphor A MATERIAL INGESTION PROCESS IS A MATERIAL CONTACT PROCESS on the effects of this action on the Actor-Goal, which usually result in a change of state.