Focus without pitch boost: focus sensitivity in Japanese why-questions and its theoretical implications

Abstract

Unlike typical wh-questions, why-questions are known to be focus-sensitive, but the linguistic realization of their focus sensitivity shows an unexpected pattern in Japanese. The phrase that immediately follows a causal wh-phrase can be considered as the focus associate without any focal prominence. This prosodic pattern contradicts the generally accepted view that a focused phrase invariably receives focal prominence (pitch boost) in Japanese. The paper presents an analysis based on focus movement for this surprising prosodic pattern. We characterize the focus sensitivity of a why-question as an association-with-focus effect with the silent focus exhaustivity operator. The adjacency of a causal wh-phrase and the focus associate is a result of the focus movement to the operator position, which mimics the focus movement proposed by some of the advocates of focus association by movement (Krifka in The Architecture of Focus 82:105, 2006; Wagner in Natural Language Semantics 14(4):297-324, 2006; as reported by Erlewine (Movement out of focus, 2014)). We argue that the adjacency strategy, which places a focus associate immediately after why, is a syntactic manifestation of association with focus, and that this structural disambiguation makes prosodic marking unnecessary. The proposal brings a functional perspective to the syntax–semantics–prosody correspondence in such a way that a focus-marked phrase does not automatically lead to prosodic prominence and the phonological interpretation of focus is influenced by the consideration of usefulness.

Appendix: Exhaustivity in Wh-questions

The lack of exhaustive implicatures in constituent questions is discussed in Tomioka (2020).

One of the key factors in the proposed analysis in the paper is that a focused phrase in a why-question is associated with the exhaustivity operator (${exh_{F}}$), rather than why itself. The fact that other constituent questions are not focus sensitive means that they cannot host $\textit{exh}_{F}$, and one naturally wonders why that is so. This appendix addresses this issue, but we can offer only a preliminary analysis.

In many information-structure-based analyses of wh-questions (e.g., Krifka 2001a), wh-phrases themselves are regarded as foci, and the question of whether there are additional (non-wh) foci in constituent questions is a non-trivial matter. On the one hand, it has been observed that a (non-why) wh-phrase and a focus cannot co-occur in some languages. Italian is a well-known case (cf. Rizzi 1997; 2001). In connection with intervention effects in wh-questions, Tomioka (2007) claims that a (non-wh) focus in a wh-question is illegitimate in Japanese, regardless of their surface positions, and Bocci et al. 2018 offer an analysis that unites Italian and Japanese in this regard. On the other hand, English wh-questions can seem to host additional focused items, as in (2), repeated as (56).

While some have argued for the possibility of a non-wh focus in a constituent question (e.g., von Stechow 1982; Kadmon 2001), the focused phrases in (56) are realized as contrastive topics in the translations into languages in which contrastive topics are easily distinguished from foci (with Japanese being one such language). Whether non-wh foci in constituent questions are unambiguously contrastive topics or not, one fact is clear: focused phrases in constituent questions cannot be interpreted exhaustively. For instance, the question, who did ANNA meet on Friday? with focal prominence on ANNA, cannot mean ‘which person x is such that it is Anna that met x on Friday?’. If this interpretation were available, the following conversation would be well-formed.

C’s utterance was meant to correct the exhaustivity associated with ANNA, but its infelicity indicates that exhaustivity is absent in A’s question.

This in turn means that the following structure is not permitted:

We propose the following descriptive constraint to block (58):

(48) can differentiate why-questions from other wh-questions. It has been suggested (e.g., Rizzi 2001; Ko 2005) that why is merged directly into the specifier of CP without leaving any trace within the TP. For instance, the two sentences in (1) have the following LF representation, in which ${exh_{F}}$ is attached to the TP that contains no wh-traces.

As for the source of the constraint in (48), we speculate that it is based on interpretability. Suppose that all wh-phrases have Hamblin-style denotations and are moreover interpreted in situ (via LF reconstruction or the deletion of the higher copy in the Copy-and-Delete theory of movement). There are at least two compositional mechanisms to implement the Hamblin semantics.

Whichever mechanism is chosen, the exhaustivity operator ${exh_{F}}$ cannot perform its duty: it is a sentential operator whose prejacent is a proposition. Its contribution to the meaning is to exclude non-weaker alternatives to its prejacent. If we adopt (61a), then, a wh-containing TP denotes a set of propositions, rather than a proposition. Thus, the exhaustification is undefined. The other choice also fails to work for ${exh_{F}}$. If ${exh_{F}}$ is attached to a constituent containing a wh-phrase but not the relevant Q-morpheme, the ordinary semantic value of that constituent is undefined. Since ${exh_{F}}$ negates all the non-weaker alternatives to the ordinary semantic value, the operation cannot be performed if the ordinary semantic value is undefined.

While the proposal explains the lack of focus-related exhaustivity in non-why constituent questions, it crucially relies on the assumption that all wh-phrases reconstruct and are interpreted in their original positions. It remains to be seen if this assumption is independently justified.