Northern Norwegian Degree Questions and the Syntax of Measurement

Norwegian dialects, including Northern Norwegian (NN), make use of degree questions with no overt degree operator (Null Degree Questions, NDQs). These questions have a gradable adjective in situ and subject-verb inversion, for example Er du gammel?, literally ‘Are you old?’, has the interpretation “How old are you?” In this paper we provide a detailed syntactic and semantic analysis for NDQs in NN which provides new insight into the decompositional semantics of adjectives more generally. We show how our analysis fits in with current research into the syntax and semantics of measure phrases and comparatives. 1. Degree questions in Northern Norwegian 1.1. Various Scandinavian degree question operators Scandinavian languages generally form degree questions in the same way as English does, where a question word (English how) pied-pipes a degree adjective. ∗To appear in Frascarelli, Mara (ed.), Phrases of Interpretation, The Hague: Mouton de Gruyter. Kristine Bentzen contributed importantly to the collection of the data used throughout this paper; thanks to her and to the other consultants, including Marit Richardsen Westergaard and Merete Anderssen. We are also grateful to the audience at IGG XXXI, and to Marit Richardsen Westergaard and an anonymous reviewer for insightful comments on earlier versions of this work. This material is based upon work supported by the National Science Foundation under Grant No. 0094263 and by the American Council of Learned Societies through a Charles A. Ryskamp Fellowship to Chris Kennedy. (1) a. How old are you? b. How far is it to Alaska? Here we illustrate with Icelandic in (2) and with Norwegian in (3); we gloss Icelandic hversu as ‘how much’ because it is only used as a degree operator, lacking the manner adverbial use of English how, and we gloss Norwegian kor as ‘where’ because it has a locative meaning in the absence of an adjective (as in Kor er han? ‘Where is he?’). (2) a. Hversu how.much gammall old ertu? are.you ‘How old are you’? b. Hversu how.much langt far er is til to Þingvalla? Thingvellir ‘How far is it to Thingvellir?’ (Icelandic) (3) a. Kor where gammel old er are du? you ‘How old are you?’ b. Kor where langt far er is det it til to Nordkapp? North.cape ‘How far it is to the North Cape?’ (Norwegian) In addition to the above strategy, Icelandic can front a different question word, corresponding approximately to English what, and leave the adjective in situ. (4) a. Hvað what ertu are.you gammall? old ‘How old are you?’ b. Hvað what er is langt far til to Þingvalla? Thingvellir ‘How far is it to Thingvellir?’ (Icelandic) Norwegian dialects in the northwest and north of Norway (stretching from Nordmore and Trondelag northwards; Endresen 1985) also have an alternative to (3), which like the Icelandic option has the adjective in situ. Unlike the Icelandic construction, there is no overt question word. Instead, as we will demonstrate in section 2, there is a phonologically null counterpart to Icelandic hvað: (5) a. Er are du you gammel? old ‘How old are you?’ b. Er is det it langt far til to Nordkapp? north.cape ‘How far is it to the North Cape?’ (Northern Norwegian) The questions in (5) are string identical to yes-no questions in these dialects, but they are intonationally distinct (see the next subsection). They are unambiguously interpreted as degree questions by speakers of these dialects, though speakers in the southeast and south of Norway do not recognize them as such and often find them confusing or mistake them for yes-no questions (as noted by Endresen 1985). In this section, we show that the degree interpretation is not a pragmatic reinterpretation of a yes-no question, and point out some restrictions on the construction. We draw our data from Northern Norway, in particular around Tromso and Narvik; henceforth we will refer to these dialects as NN.1 There is some small variation in the way the construction is used from one dialect to the next (see Bull 1987 for some examples), but to the best of our knowledge our remarks hold for dialects as far south as Nordmore, and for the Icelandic construction, apart from the fact that the Icelandic counterpart of the operator has phonological content. We will suggest, then, that degree questions can vary at least along two parameters: whether the degree question operator pied-pipes the adjective with which it is associated, and whether the degree question operator is pronounced. We have found in Scandinavian languages three of the four possible combinations of these two points of variation. It is an open question whether a language could also have a null degree operator which pied-piped an adjective, yielding degree questions roughly like “Old are you?”2 1Except where noted, all examples in this paper are NN, in a standardized orthography, lightly modified to reflect salient dialectal features. In some cases we have based examples on Endresen (1985) or Midtgard (1995) but standardized the orthography. 2In fact, such a structure is attested in Spanish in a related construction: the comparative clause of ‘comparative subdeletion’ structures (see Rivero 1981; Kennedy 2002): (i) a. La the mesa table es is mas more large long que than ancha wide es is la the puerta. door ‘The table is longer than the door is wide.’ b. *La the mesa table es is mas more large long que than la the puerta door es is ancha. wide 1.2. Syntax/Semantics versus Pragmatics Previous work on what we will call Null Degree Questions, or NDQs, in Norwegian dialects has concurred that they cannot be analyzed in terms of a pragmatic interpretation of a formal yes-no question (Endresen 1985; Bull 1987; Midtgard 1995), and we agree with this assessment. First, as noted above, the construction is systematically intonationally distinct from a yes-no question. Yes-no questions tend to have their prosodic peak on the most deeply embedded part of the sentence, which may be a predicative adjective, for example cold in Are you cold? NDQs never have an intonational peak on the adjective, but typically have the stress further to the left (Endresen 1985). Here we use underlining to represent intonation peak. (6) a. Vil will du you ha have mange? many ‘Do you want many?’ b. Vil will du you ha have mange? many ‘How many would you like?’ (7) a. Har have du you vaert been her here lenge? long ‘Have you been here long?’ b. Har have du you vaert been her here lenge? long ‘How long have you been here?’ As Endresen notes, the yes-no questions above (the first member of each pair) may pragmatically elicit more informative responses than a simple yes or no, depending on context. But NN speakers interpret the NDQs (the second member of each pair) as unambiguously asking for degrees. This leads to a sharp difference of intuition with southern speakers (and English speakers, for the translations). Unless southern speakers are claimed to have a different pragmatics, this suggests that the difference is grammatical, not pragmatic. A second argument that the construction is not simply pragmatic comes from the difference between the presuppositions of degree questions and yesIt is well established that a comparative clause is an A-movement structure, so the existence of comparatives like (ia) suggests that degree questions with the same form should indeed be possible. no questions involving adjectival predications. A yes-no question is based on the semantics of the ‘positive’ (morphologically unmarked) form, which involves reference to a contextual standard of comparison: Is he old? asks whether an object’s age is greater than the prevailing standard of age (for things like X ) or not, and crucially presupposes that the answer to that question is not part of the common ground. In contrast, a degree question asks for the degree to which an object possesses some property: How old is X asks for an object’s age, and presupposes that this value is unknown, but it is completely indifferent as to whether X is old or not. The prediction is that in a context in which it is part of the common ground that X is old, only a degree question will be felicitous. This is illustrated by the English translations in the NN discourse in (8); the felicity of B’s question shows that it must have the semantics of a degree question. (8) A: Laereren the.teacher min my er is skikkelig properly gammel. old ‘My teacher is really old.’ B: Er is han he gammel? old ‘How old is he?’ ‘??Is he old?’ A third indication that the degree interpretation is not simply pragmatic is that various degree modifiers can be used in NDQs which are infelicitous in yes-no questions, for example sann cirka, ‘approximately,’ as pointed out by Midtgard (1995). (9) a. Er is den it tung, heavy sann such cirka? approximately ‘How heavy is it, approximately?’ b. *Er is den it tung,