ConSOLE

Comprehension and Acceptability of the English Resultative and Depictive Constructions by Spanish Native Speakers

ConSOLE29

Alejandra Celi CONICET-Argentina

English Resultative Construction (ERC) as a key instance of the English style in the design of event representations that differs considerably from the Spanish style in event construction and this, in turn, has a significant impact on the acquisition of English by Spanish native speakers (Alonso, 2015).

To what extent ERC are acquired by Spanish speakers and how subjects access English event representation, if that is the case.

Our tour

01. ERC

03. Method

04. Results

02. Objectives

.01

Antecedents

English-Spanish acquisition of the lexicalization of manner of motion and path in boundary crossing events (Cadierno, 2004, Cadieron & Ruiz, 2006, Robinson, et al, 2009, Larrañaga, et al., 2006, Alonso 2015).

The acquisition of ERC by different languages such as Chinese (Cao & Zhou’s, 2014), Portuguese (Oliveira, 2016) and Korean (Kim, et al, 2019).

L1 has an effect on the acquisition of L2, which seems to be modulated by how distant languages are in terms of Talmy's typology of motion events (Talmy, 2000) and how different the structures are.

.02

What is interesting about the Spanish-English contrast?

English is an instance of a Satellite-framed language (Talmy, 2000) whose more frequent strategy to build a motion event is by means of a main verb that tends to codify manner.

If result needs to be codified, English resorts to the Resultative Construction (ERC), a complex predicate structure with a main verb that codifies manner and a satellite phrase (either a Prepositional Phrase or an Adjectival Phrase) codifies result

Mary danced John tired.

In ERC, manner and result are causally related and they form a complex verb [CAUSE ([Mary-dance-John (e5')], [John-BECOME-tired (e5'')], (e5))]

Spanish

Spanish is an instance of Verb-framed languages in which the main verb tends to codify result and manner (if required) is lexicalized in an adjunct

Depictive Construction

A double predicate construction that is syntactically similar to the ERC (ERC-Property). Yet, semantically, it lacks the result meaning of the ERC.

[Spa] Él devolvió el libro dañado [Eng] He returned the book damaged

[Spa] María cansó a Juan bailando [Eng] *Mary tired John dancing

But what about ERC?

They form family of constructions (Goldberg, 2004) that behave as a complex verb (Wechsler, 2005) and that follows the Direct Object Restriction (DOR) (Levin y Rappaport , 1995).

ERC-pROPERTY

ERC-Fake Reflexive

ERC-pATH

change of location

change of state

[E] The baby cried himself asleep *The bady cried himself [S]'El bebé se durmió llorando'

[E] Bill rolled the ball up the hill [S] 'Bill subió la pelota por la colina rodando'

[E] Mary danced John tired. *Mary danced John[S]'Mary cansó a John bailando'

Direct Object Restriction (DOR) states that the result of a ERC is predicated fromthe argument that corresponds to the DO (ainstead of the Subject). The ERC with intransitive verbs can introduce a non-lexically licensed DO which in turn controls the predicative phrase.

Strong ERC (Washio, 1997)

ERC that encompass a change of state or location that is not entailed by the main verb meaning

Mary danced John tired.

Weak ERC

Spanish can only express those where the result is entailed by the main verb meaning while it cannot express those weak ERCs where the verb just implicates it (Paris, 2019)

The result is either lexically entailed or strongly implicated.

[Eng] My neighbor painted his house green. [Spa] Mi vecino pintó la casa verde.

[Eng] Mary wiped the table clean [Spa] *Mary repasó la mesa limpia Mary limpió la mesa

.03

Objetives and Predictions

if this sequence occurs, to identify if it is modulated by their departure from the Spanish pattern and proficiency level

To analyze if DC are acquired at an earlier stage than ERC

To identify if the different subtypes of ERC are acquired in a given sequence

- DC are acquired at an earlier stage than ERC since they belong to the Spanish style in event construction -Proficiency level would improve the acquisition of ERC in general, but this effect would interact with type of ERC. - Constructions that are closer to Spanish, like DC, would be easier to acquire even for low proficiency speakers when compared to ERCs that do not have a parallel construction in Spanish. Therefore, these ERCs that are more distant from Spanish would be more subjected to proficiency effects.

.04

Methodology

Task 2: Acceptability Judgement Task

Task 1: Sentence Comprehension

INFO

INFO

• online questionnaire
• self reported proficiency questionnaire
• vocabulary LexTALE Test (Lemhöfer, K & Broersma, M., 2011)
• frequency of use and immersion in L2 questionnaire Language History Test (Li, Zhang, Tsai & Puls, 2014).

Task 1: Sentence Comprehension

"Read the following sentences and choose the option that better describes their meaning. If you don't know the answer, chose "d" Tom watered the plants flat a) The plants became flat because Tom watered them b) The plants were already flat when Tom watered them c) Tom was feeling flat when he watered the plants d) Do not know/ Do not answer

• 285 Spanish native speakers
• Different levels of proficiency in EFL (low, intermediate and hight) accoring to LexTALE scores (n = 285, mean = 46.8, sd = 6.98, range: 28 to 60)
• 36 stimuli
• 24 target sentences: (6 Property ERC, Path ERC, Fake Reflexive ERC, and DC) + 12 distractors
• 3 possible interpretations (plausible and correct, plausible but incorrect, implausible and incorrect)

Results

TASK 1: SENTENCE COMPREHENSION

A mixed regression model with accuracy of response (correct/incorrect) was carried out to evaluate the interaction between Sentence Type and Proficiency in L2, which turned out significant (χ2 (3) = 68.246, p < 0.0001).

• Responce accuracy increases as LexTALE scores increase in all sentences type (z’s < -3.360, p’s < 0.0008).

• The performance in the comprehension of DC is higher than Property ERC (z’s > 3.58, p’s < 0.001) and Fake Reflexive ERC (z = 3.89, p = 0.0006) in the lower LexTALE level, which is not observable in intermediate and high level

• DC do not differ significantly fron Path ERC at all LexTALE levels INCLUDE STADISTICS

Back

Task 2: Acceptability Judgement Task

• AJT Likert scale de 1-7
• 2 forms
• 90 subjects with different proficiency level in EFL (low, intermediate, high) (mean = 46.65, sd = 7.49, range: 29 to 60)
• 64 stimuli
• 48 target sentences in 3 conditions: (Property ERC, Fake reflexive ERC, DC) + fillers
• 16 sentences per condition (8 acceptable + 8 with semantic violations)

“How acceptable would the following sentences sound to a native speaker of English? Please respond as fast as you can.”

ERC of Property Robert danced his feet sore at the party *Robert danced his feet painful at the party ERC of Fake Reflexive We yelled ourselves hoarse at the football match *We yelled ourselves loud at the football match DC object oriented John returned the book damaged yesterday *John returned the book old yesterday

Mixed –effects linear regression model as a continuous predictor were conducted on correct and incorrect items separately. Both models turned out significant (χ2 (2)’s > 8.340, p’s < 0.015).

TASK 2: AJT

• All three levels of vocabulary proficiency (35, 45 and 55) displayed the same pattern: higher acceptability ratings for DC compared to both ERC Property and Fake Reflexive (T’s > 4.578, p’s < 0.001), with no significant differences between the latter

• Acceptability increased with LexTALE scores for Property (t = -2.604, p = 0.027) and Fake Reflexive (t = -3.424, p < 0.002) resultatives, but not for DC (t = -1.06, p = 0.5).

• Acceptability of unacceptable DC (Estimate: 0.526, std error: 0.193, T = 2.733, p = 0.019) and Property ERC (Estimate: 0.462, std error: 0.193, T = 2.398, p = 0.047) decreases with LexTALE scores
• while no significant changes are observed for Fake Reflexives.

Back

.06

1) The four conditions are comprehended differently being the DC the best comprehended and accepted by Spanish speakers for all LexTALE scores.

2) The comprehension and acceptability of ERC improve with vocabulary proficiency (DC/Path ERC > Property ERC > Fake Reflexive ERC).

Conclusions

This may be due to the effect of the event construction of L1 on the comprehension of English as L2 since Spanish speakers can use their knowledge about DC even in low proficiency levels as they function as a mirror image of one another

It seems that there is a sequence of acquisition of the different ERC. This effect may be driven by the departure of each structure from the Spanish pattern of event construction.

ERCs that are more distant from Spanish would be more subjected to proficiency effects

Referencias: Goldberg, A & Jackendoff R. (2004) English Resultatives as a Family of Constructions. Language, 80, (532-569.) Lemhöfer, K & Broersma, M. (2011) Introducing LexTALE: A quick and valid Lexical Test for Advanced Learners of English. Springerlink.com disponible en https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3356522/ LEVIN, B., & RAPPAPORT HOVAV, M. (1995). Unaccusativity: At the syntax-lexical semantics interface. Cambridge, MA: MIT Press.Li, P., Zhang, F., Tsai, E., Puls, B. (2014). Language History Questionnaire (LHQ 2.0): A New Dynamic Web-Based Research Tool. En Bilingualism: Language and Cognition, 17(3), 673-680. DOI: 10.1017/S1366728913000606 Paris, L. (2019). La construcción resultativa del inglés en contraste. Onomázien 46: 259-286 DOI: 10.7764/onomazein.46.13 ISSN: 0718-5758 RAPPAPORT HOVAV, M., & Levin, B. (2001). An Event Structure Account of English Resultatives. Language 77, 766-797. Talmy, Leonard (2000). Toward a cognitive semantics, vol. 2: Typology and process in concept structuring. Cambridge, Mass.: MIT Press. Washio, R. (1997) Resultatives, compositionality, and language variation. Journal of East Asian Linguistics 6, 1–49 Wechsler S. (2005) Resultatives under the Event Argument Homomorphism in The Syntax of Aspect, ed. Nomi Erteschik-Shir and Tova Rapoport, Oxford University Press

Thank you for your time!

Any questions/suggestions/comments?

maceli@mendoza-conicet.gob.ar

.05

We ran a series of linear mixed-effects regression models on correct and incorrect items separately. Both models turned out significant (χ2 (2)’s > 8.340, p’s < 0.015).

What we found

Acceptable conditions

• All three levels of vocabulary proficiency (35, 45 and 55) displayed the same pattern:
• higher acceptability ratings for DC compared to both ERC Property and Fake Reflexive (T’s > 4.578, p’s < 0.001),
• with no significant differences between the latter

• Acceptability increased with LexTALE scores for Property (t = -2.604, p = 0.027) and Fake Reflexive (t = -3.424, p < 0.002) resultatives, but not for DC (t = -1.06, p = 0.5).

Table 1: Post-hoc analyses Lextale per Construction Type in grammatical sentences; p-values adjusted by Tukey method.

In the analysis of items with semantic violations, the nested model comparison showed significant improvement of fit after adding the interaction (χ2(2) = 17.359, p < 0.001).

Non acceptable sentences

• Acceptability ratings for DC and Property ERC with anomalies were higher than those of Fake Reflexive at low LexTALE scores, while this pattern was reversed at high LexTALE scores.
• No significant differences were observed at intermediate LexTALE scores.

Table 2: Post-hoc analyses Lextale per Construction Type in ungrammatical sentences; p-values adjusted by Tukey method.aquí