Stable and developmental optionality in native and non-native Hungarian grammars

Abstract

Listen

Translate article

The existence of optionality is well attested in natural languages. It can be defined as the coexistence of two variants of a given construction with identical Logical Form (LF) representations within the same grammatical system. The existence of optionality within both stable (native and non-native) and developing grammars seems to contradict the fundamental claim of the Minimalist Program (MP). In the MP, syntactic optionality is excluded within the computational system because of economy principles which require an ‘optimal realization of interface conditions' (Chomsky, 1995: 171).As a result, movement is possible only if it is necessary. Movement which is possible but not necessary (i.e., truly optional) is not permitted. Thus, at the level of competence, optionality may occur precisely because neither alternative qualifies as a unique optimal realization of interface conditions.This article presents evidence from the grammaticality intuitions of native speakers and English-speaking second language (L2) learners of Hungarian about optional and categorical constructions. It is argued that while often observed, stable native optionality is highly restricted and difficult to capture. This might be due to the fact that developmental L1 optionality disappears when the input that children receive is sufficient and/or sufficiently robust. Where the input is not sufficient, optionality lingers on, creating fuzzy grammars and borderline cases in grammaticality. However, such stable native optionality may eventually be resolved by categorization (one alternative is expunged from the grammar) or by semantic differentiation (leading to quasi-optionality).While stable native optionality is unpredictable, developmental optionality is relatively predictable in both early and end-state non-native grammars. The nature of early L2 optionality depends on the characteristics of the initial state. Advanced and end-state non-native optionality persists only if the target language input is non-robust or parametrically ambiguous. Therefore the resolution of optionality at ultimate attainment depends not only on the L2 initial state but also on the intrinsic nature of the primary linguistic data that the learners are exposed to. It is well known that the evidence L2 learners receive is qualitatively different from that obtained by first language (L1) acquirers, and quantitatively limited compared to it. This, in conjunction with the hypotheses created initially, may lead to non-native-like representations that are nonoptimal realizations of the interface conditions. In the face of non-robust and parametrically ambiguous data, L2 learners are either found to fall back on L1 values or resort to general learning strategies, thus ultimately exhibiting divergent representations in end-state grammars.