RRL, LXII, 1, p. 59–76, Bucureşti, 2017
PROSODIC EFFECTS
*ION GIURGEA1
Abstract: I argue that a certain intonational pattern found in Romanian polar questions, which has been qualified as neutral in previous studies, actually marks verum focus. This pattern consists of a L* nuclear accent on the verb and a HL% boundary tone, with the H part associated to the last stressed syllable, and the L part truncated when this syllable is the very last in the sentence. This pattern also characterizes early constituent focus (including focus on the verb, which is realized in the same way as verum focus). The neutral intonation is different, consisting of a L* nuclear accent on the last stressed constituent followed by a H% tone. This pattern is identical to the one found with narrow focus on the last constituent, as expected given the general prosodic properties of the language.
Keywords: verum, focus, polar questions, intonation.
1. INTRODUCTION: VERUM FOCUS AND QUESTIONS
The term verum focus, introduced by Höhle (1988, 1992), refers to a type of focalization manifested by sentence stress on an element that fills a (dedicated) clausal functional head position, C or Infl (which extends to the Spec of a null C, for German2), which, for declaratives, can be roughly characterized as emphasizing the assertion:
(1) a. But I DID take care
b. Dar am AVUT grijă (Ro.) but have.1SG had care
c. Ich HABE doch aufgepasst (Ge.) I have.1SG but taken-care
* This work was supported by a grant of the Romanian National Authority for Scientific Research and Innovation, CNCS – UEFISCDI, project number PN-II-RU-TE-2014-4-0372.
1 The “Iorgu Iordan – Alexandru Rosetti” Institute of Linguistics of the Romanian Academy, Bucharest, [email protected].
2 In German embedded interrogatives, verum focus can be realized by stress on the wh-pronoun that occupies SpecCP.
In English, where this focus characterizes Infl and can trigger do-support (see (1)a), it has been described as focus on polarity (Halliday 1967, Gussenhoven 1984) or on the truth value of the assertion (Dik et al. 1981). Höhle’s own characterization, as focus on a truth predicate VERUM, is quite close to the latter view. The semantic characterization of this phenomenon is still controversial (see Lohnstein 2016 for an overview of the various proposals): there is disagreement on whether it represents focus on an element of meaning present in other sentences (as in Höhle’s original proposal, or in the polarity focus analysis) or a special operator (Romero and Han 2004, Romero 2005), which may even not involve focus at all – Gutzmann and Castroviejo Miró (2011), Repp (2013), Gutzmann et al. (2017) defend the view that verum is a conversational operator independent from focus.
Among those who include this phenomenon under focus, there is no consensus about the element that bears the focus feature: besides polarity and a special operator verum, this element has been identified with “sentence mood” by Lohnstein (2012, 2016), a notion roughly corresponding to sentential force – assertive, interrogative, imperative – but not limited to main clauses.
There is no substantial disagreement about the pragmatic conditions of the phenomenon: the propositional content p must be contextually given (see Büring 2006), sometimes with a different polar component – e.g., (1) can be a reply to
‘But you probably didn’t take care’. The antecedent can also differ in modality and the illocutionary component. Here are other illustrations of the contexts that elicit verum focus, which also show that verum focus can be found in all sentence types – declarative, interrogative, imperative (in (2) and (4), as well as in (1) above, I used light verbs, in idioms – avea voie ‘be allowed’, da voie ‘give permission’, avea grijă ‘take care’ – and in (3) a copula, in order to show that verum focus is prosodically marked on the inflected verb, as opposed to predicate focus, which involves nuclear stress on the lexical predicate or on the last member of the idiom;
note that in Romanian auxiliaries behave as clitics – see Guţu-Romalo 1962, Dobrovie-Sorin 1994 – so the nuclear stress on Infl is phonologically realized as nuclear stress on the lexical verb also when Infl is an auxiliary – see (1)b)3:
(2) [Context: people are debating on whether somebody’s actions were legal or not]
AVEA voie had permission
‘He WAS allowed to do that.’
(3) [A: Ne aşteptam toţi să fie supărată ‘We all expected her to be upset.’]
B: Şi, ERA supărată?
and was.3SG upset
‘So, WAS she upset?’
3 Nuclear stress on the inflected verb can also be used to mark narrow focus on Tense (cf.
Zimmermann 2016). In the examples discussed here, this is clearly not the case.
(4) [Context: X said he was unsure whether he should allow his daughter to play outside; addressed to X:]
DĂ-i voie!
give.IMPV-3SG.DAT permission
‘DO allow her!’
Höhle also includes here instances of focus in embedded contexts, where the antecedent contrasts in polarity with the clause at-hand; here is a Romanian example modeled after Höhle’s (1992: 134):
(5) Mi-ai spus doar ce n-am voie. Acum aş vrea să me-have.2SG told only what not-have.1SG permission now would.1SG want SBJV ştiu ce AM voie să fac
know.1SG what have.1SG permission SBJV do.1SG
‘You only told me what I’m not allowed to. Now I’d like to know what I AM allowed to do.’
The analysis as focus on polarity works for contrasts in embedded clauses as in (5) and can be extended to declaratives in general and even imperatives, but it is hard to see how it can deal with interrogatives as in (3). Since we are interested here in verum focus in questions, I will distinguish verum focus from focus on polarity, excluding from the discussion examples of the type in (5). The fact that (3) does not represent focus on polarity becomes clear if we consider its negative version. A negative counterpart of (3) cannot use focus on a preverbal negation in the same way (i.e., to mark verum focus): English may resort to a low negation below Infl, and verum focus remains on Infl ((6)a); in Romanian, where this option is impossible because sentential negation is always a functional head above Infl4, no negative counterpart of (3) can be built; (6)b, with focal stress on Neg, has a different interpretation, that of a positively biased question, as its English counterpart with high Neg (WASn’t she upset?). (6)c is also ruled out in this context, presumably because VERUM must be interpreted above Neg (cf. (6)a), but the auxiliary must stay below, so the verum focal stress has no element to attach to:
(6) [A: ‘We all expected her not to be upset.’ – Ro.: Ne aşteptam toţi să nu fie supărată]
a. B: ‘So, WAS she not upset?’
b. # Şi, NU era supărată? (Ro.) and not was.3SG upset
c. # Şi, nu ERA supărată?
and not was.3SG upset
4 More precisely, in the complex-head analysis of the Romanian inflectional complex (formed by V and various clitics, including auxiliaries; see Dobrovie-Sorin 1994, Barbu 1999, Dobrovie-Sorin and Galves 2001, Giurgea 2011), Neg is part of this complex, and is situated above what we may call
‘Infl’, the position where mood, tense and subject agreement are realized.
For questions as well as assertions, the condition for the use of verum appears to be that the issue ‘whether p’ should be contextually active (be part of the background of the conversation, with a certain degree of activation). For declaratives, not only questions but also assertions about p make good antecedents, because any assertion raises the issue of its acceptance by the hearer, before entering the common ground. For commands, as in (4), the antecedent must be a modalized proposition, referring to a possible future course of action of the addressee (the open issue is whether p should be added to the addressee’s to-do list).
The fact that the propositional content must be discourse-given, alongside the use of nuclear stress, with deaccentuation or substantial compression of following accents, support the view that the verum phenomenon does indeed involve focus.
As for the alternatives involved, the set {p, ¬p} works for assertions and commands, but it is not clear how it can apply to questions.
In the case of questions, as we have seen, the condition is that the issue
‘whether p’ be already present in the discourse context – in (3), it is inferable from the previous discourse; but it can also be a question previously raised and left unsolved:
(7) [context: B wants to hear from A whether C solved a problem; A tells B various things about what C did, but it is still not clear whether the problem was solved]
B: N-am înţeles, până la urmă: A REZOLVAT problema?
not-have.1SG understood in-the-end has solved problem-the ‘I can’t see, in the end: DID he solve the problem?’
Here, the propositional content p is given, and even the illocutionary force of the antecedent is the same. However, as the addressee behaves as if the question
‘whether p’ has been solved, focus is used to highlight the fact that this is not the case, that ?p is still among the questions-under-discussion. We can thus say that the illocutionary component represents the focal part, as predicted by Lohnstein’s (2012, 2016) analysis.
Besides this type of use, Giurgea and Remberger (2012, 2014) claimed that verum focus in polar questions can also have another interpretation in Romanian, that of a mirative focus, in which the speaker presents the proposition at hand as surprising, less likely w.r.t. its alternative(s), which yields a biased question interpretation – the speaker expects an opposite answer (for a formal characterization of mirative focus, see Bianchi et al. 2015, 2016). However, a closer look at prosody shows that the two types of focus must be kept apart.
Giurgea and Remberger discuss sentences in which the normal declarative order has a preverbal subject (because the type of predicate does not allow a thetic construal) but, when used as polar questions, display VSX orders correlated with emphasis on the verb. A case in point is (8). Here, VSX can be used to reactivate an unsolved question, as in (8)a, what I have called ‘verum focus’ so far, but also to express surprise, with a mirative interpretation, as in (8)b. However, as shown in
the two diagrams (fig. 1, for (8)a, and fig. 2, for (8)b), the two sentences have a different prosody5:
(8) a. N-am înţeles până la urmă: ARE fratele tău maşină?
not-have.1SG understood in-the-end has brother-the your car H+L* H* L%
‘I didn’t understand, in the end: DOES your brother have a car?’
b. E nu zău. ARE fratele tău MAŞINĂ?
really has brother-the your car
L* H% L+H* L+H* !H%
‘Really? Does your brother have a car?’
Figure 1 (example 8a).
(8)a clearly has nuclear stress on the verb, manifested by a low tone, as is typical for focus in polar questions (cf. Avram 1973, Dascălu 1979a, Grice et al.
2000, Dascălu-Jinga 2001). The final intonation is also typical for polar questions with early focus, as we shall see in the next section: a HL% tone realized as H* on the last accented syllable followed by a fall in the last syllable (on the view that the H* L% sequence represents the secondary association of an HL boundary tone with the last lexical accent, see Grice et al. 2000, Ladd 2008).
5 The prosodic annotations follow the conventions of the ToBI framework (cf. Pierrehumbert 1980, Pierrehumbert and Hirschberg 1990, Ladd 2008): H = high tone, L = low tone, * = tone associated with the lexically accented syllable, % = boundary tone of an intonational phrase, – = boundary tone of an intermediate phrase.
Figure 2 (example 8b).
(8)b, on the other hand, shows a rising accent on the verb (L*+H), which might result from a L* typical of focus in polar questions followed by a H- intermediate boundary tone, and a rise on the last accent, followed by a fall to a mid level on the final, unaccented syllable (which I analyze as L+H* !H%); as this pattern is different from the one characteristic of polar questions with early focus (see next section), I treat it as nuclear accent. Notice also that a H* tone appears on the last accented syllable of the intermediate constituent fratele tău ‘your brother’, which is characteristic of prenuclear accents (fratele tău is equally given in (8)a and (8)b; the reason it is treated differently must be related to its position relative to the nuclear accent; it is well-known that deaccenting – a level low tone or a very reduced pitch movement – is manifested in postnuclear accents, cf. Halliday 1967, Büring 2007, 2012, Ladd 2008, Rochemont 2013, 2016, a.o.).
This difference in prosody can be partly due to the fact that mirative focus involves a special prosody (as has been found for Italian by Bianchi et al. 2015, 2016), but in view of the final tonal configuration, it is not clear yet that we are dealing with an early focus, therefore I will not include the type in (8)b in the discussion that follows.
Even if (8)b has an early focus, it must lie on a distinct element than in (8)a:
in (8)b, the alternatives are {p, ¬p}: as p is less expected than ¬p, we derive the incredulity meaning. The question operator is above focus. In (8)a, as I have argued, it is the question operator itself which bears narrow focus6.
6 Negative questions expressing doubt (i.e., with a positive bias, the negative counterpart of (8)b) do have the stress pattern of early focus, but with nuclear stress on negation. This can be analyzed as mirative focus on polarity (I use an example where the verb does not begin in a vowel, so that the stress on negation can be seen clearly):
(i) NU va veni fratele tău cu maşina?
not will.3SG come brother-the your with car-the H+L* H* L%
Summing up, we can conclude that
(9) In polar questions, verum focus signals that the question q is already among the questions under discussion. The focused element is the question operator (the illocutionary component of the sentence).
2. THE PROSODY OF POLAR QUESTIONS WITH AND WITHOUT VERUM FOCUS
We have seen in (8)a that in order to raise again a question that has not yet been solved, a verum focus pattern can be used, characterized, in Romanian, by a low nuclear accent and a HL boundary tone which realizes its H part on the last lexical accent.
In the following, I will argue that this tonal pattern has been misinterpreted as the neutral intonation of polar questions in Romanian. I will show that the neutral intonation of polar questions does not involve a peculiar early nucleus on the verb, but has the nuclear stress on the last (non-deaccented) lexical accent, as in other Romance languages and in Germanic. The cases of early nucleus on the verb can be analyzed as instances of verum focus, under the characterization in (9). This will also shed further light on the pragmatic conditions of the verum focus phenomenon in questions.
I will restrict the prosodic discussion to standard Romanian. It is known that some regional varieties, such as those of Muscel, Bihor, Maramureş, Northern Transylvania, have different prosodic patterns for polar questions (Avram 1973, Dascălu 1975, 1986, Dascălu-Jinga 2001, Jitcă et al. 2015); how these varieties encode early focus in questions, including verum focus, is an issue left open for further research.
The fact that early focus in polar questions in standard Romanian is realized by an L* nuclear accent and a HL% edge tone with the H component on the last lexical accent is well-established (cf. Dascălu 1979a, Dascălu-Jinga 2001). Here is an example displaying focus fronting7:
(10) [Context: A meeting place had been established, X is no longer sure that he remembers the place correctly and checks whether the meeting place was Capşa]:
La 'CAP.şa a fost în.tâl.'ni.rea?
at Capşa has been meeting-the L* H* L%
‘Is it at Capşa that we had to meet?’
‘Won’t your brother come by car?’
7 On the interpretation of focus fronting in polar questions, see Giurgea (2016) for Romanian, Bianchi and Cruschina (2016) for Italian.
Figure 3
This pattern is very similar to the one used for verum focus in (8)a (see figure 1); the slight difference, the lowering on the nuclear syllable in (8)a, which I notated as H+L*, is obviously due to the fact that the nuclear syllable is sentence- initial in (8)a, as opposed to (10), so that the L* accent must be realized by a slight fall in (8)a, whereas in (10) the L tone is clearly signaled by a lower pitch level than on the pretonic syllable. No wonder then that the verum pattern, which had been taken as a neutral pattern by Dascălu (1979a,b), Dascălu-Jinga (2001), has been described by this author as involving emphatic stress on V, which led her to conclude that “emphasis, in questions, is a constitutive element, whereas in declaratives it is only optional” [“emfaza este un element constitutiv în întrebări, în timp ce, în enunţurile declarative, ea este numai opţională”] (Dascălu-Jinga 2001: 33).
Besides Dascălu-Jinga, the pattern with nuclear stress on V and an HL at the end has been considered neutral by Grice et al. (2000), Ladd (2008).
I will show that this pattern is not neutral, but involves special pragmatic conditions which can be characterized in terms of verum focus, under the formulation given in (9). The neutral pattern, used in totally out-of-the-blue questions, has the nuclear stress on the last lexical accent, as expected, realized as L* followed by a H% boundary tone. Thus, a totally out-of-the-blue question will have the following pattern:
(11) [Ce s-a întâmplat,] a ve.'nit 'ta.ta?
what REFL-has happened has come father-the L+H* L* H%
‘What happened? Has dad come?’
Figure 4
In the following, I will refer to this pattern as the ‘LH pattern’ (fig. 4), and to the verum focus one in (8)a, with a L* nuclear stress on the verb and HL in the end (fig. 1), as the ‘L..HL pattern’.
In their brief survey of Romanian intonation, Jitcă et al. (2015: 299–300) find both patterns for information-seeking polar questions:
(12) [Context: somebody enters a shop]
Tested sentence: Aveţi marmeladă? ‘Do you have marmalade?’
a. A.'veţi mar.me.'la.dă? (L.HL pattern) L* L+¡H* L%
b. A.'veţi mar.me.'la.dă? (HL pattern) L* L* H%
Although these are out-of-the-blue questions, the context is not innocuous: in the customer-seller linguistic exchange, there is a list of issues expected to occur:
whether the seller has a number of items – which the customer wants to buy – and at what price. Although these issues have not been explicitly introduced in previous discourse, maybe they can be treated as activated in the context, which would allow the verum focus intonation, i.e., the L..HL pattern in (12)a. Another possible explanation, which will be presented in section 3, is that this is a special pattern used for requests with an interrogative form.
In (11), the use of the L..HL is out of the question. The introductory clause
‘What happened?’ rules out any connection with a previously opened issue, and the sentence represents a genuine information-seeking question, not a request.
If we eliminate the ‘what happened?’ part, we can use the L..HL pattern on the sentence A venit tata? ‘Has dad come?’, but only under specific contextual conditions: father’s arrival must have occurred as an issue in the communicative exchanges between the conversation participants at some point before, although not necessarily in the immediately previous dialogue – the sentence can start a new conversation, but requires that the issue of the father’s arrival should have recently been touched to and should still be relevant – e.g., three hours before, somebody said that we have to wait for father to get home in order to watch a DVD.
I tested these intuitions on pairs of examples which contain the same polar question in different (imagined) contexts, one that allows verum focus and one that does not. Here are the sentences and the predictions:
(13) [Context: you meet some acquaintances on the street by chance]
Ce faceţi? Aţi fost la cumpărături?
what do.2PL have.2PL been at shoppings
‘What’s up? Have you been shopping?’
Prediction: no verum focus, LH
(14) [Context: some people were supposed to do the shopping; you check whether they fulfilled their commitment]
Gata? Aţi fost la cumpărături?
ready have.2PL been at shoppings
‘Ready? Have you been shopping?’
Prediction: verum focus, L..H (= L*..HL% with truncation, see discussion below) (15) [Context: you enter a room where somebody has just finished a phone conversation]
Ce faci? Ai vorbit cu copiii?
what do.2SG have.2SG talked with children-the
‘What’s up? Have you spoken with the children?’
Prediction: no verum focus, LH
(16) [Context: somebody told you a few hours ago that she would try phone her children, who are abroad]
E, ai reuşit? Ai vorbit cu copiii?
hey have.2SG succeeded have.2SG talked with children-the
‘Well now, have you succeeded? Have you spoken with the children?’
Prediction: verum focus, L..HL
I used 13 informants, all living in Bucharest, aged between 29 and 69. They were asked to read aloud the sentence, as they would pronounce it in the indicated context. The experiment contained other sentences besides these, with different prosodic phenomena. The recordings have been analyzed with Praat (Boersma and Weenink 2012).
Before presenting the results, we should notice that the sentence in (13)-(14) differs from the examples examined so far ((8), (10)-(12)), as well as from the one in (15)-(16), in that the last syllable is lexically accented – see the prosodic structure of the two sentences:
(17) a. (13)-(14): Aţi 'fost la cum.pă.ră.'turi b. (15)-(16) Ai vor.'bit cu co.'pi.i
As noticed in previous studies (Grice et al. 2000, Ladd 2008), the HL boundary tone of polar questions with early focus is truncated in this case, leaving only the H element. This results in a L+H or H tone on the final syllable. So, instead of the L..HL pattern, we will have an L..H pattern (as indicated in the prediction for (14) above). As for the tone used in neutral contexts, and, more generally, when the nucleus is last, it will display the nuclear L* tone and the boundary H% tone on the same syllable, leading to a rising intonation. So the LH pattern is preserved, but compressed on the final syllable.
Given the truncation of the HL boundary tone, the early and late nucleus patterns (verum vs. neutral, in my account) in (13)–(14) are hard to distinguish by the final configuration – in the L..H pattern (early nucleus), the pitch is low until the last syllable and then it rises in order to attain the H target of the boundary tone (which, by secondary association, can also be taken to include a H*); in the late nucleus pattern, the nuclear stress is L* but the pitch has to rise in order to realize the H% tone on the same (last) syllable, yielding the same rising effect. The two patterns are clearly distinguishable in the initial portion of the clause: in the late nucleus pattern, the first accent (on fost) is rising (as normal for prenuclear accents, see Jitcă et al. 2015): the pitch starts to rise from the onset of the syllable and reaches a maximum in the posttonal syllable, for most speakers; this can be represented as L+<H*, following the conventions in Frota and Prieto (2015); in just one case I found no higher pitch on the posttonal syllable (see pattern A´). In the early nucleus pattern, the pitch on fost is lower than in the preceding syllable and stays low until the final rise. The two schemas are summarized below:
(18) a. Aţi 'fost la cum.pă.ră.'turi?
L+<H* L*H% Pattern A L+H* L*H% Pattern A´
b. Aţi 'fost la cum.pă.ră.'turi?
L* L+H* H% Pattern B
As can be seen in table 1, all informants used pattern A for (13), as predicted, whereas for (14) only 6 used pattern B, and the remaining 7 used again pattern A/A´. This is expected if the LH pattern, i.e. late nucleus, represents the unmarked pattern, as I propose: some informants did not take into account the possibility of an open issue ‘whether p’ suggested by the indications about the context, so they used the neutral pattern.
In the second pair of examples, the same contrast in the first part of the sentence appears, but also the last accents are very different: in the LH pattern (late nucleus), the penultimate syllable, which carries the nuclear accent, has a low tone (L*), possibly with a fall (H+L*; see pattern C´), and then the tone rises
considerably on the last syllable, to realize the H% tone. In the L..HL pattern (early nucleus), the penultimate syllable must bear the H* tone, which is normally realized by rising from the low plateau that follows the L* nucleus (pattern D), but can also involve a rise on the preceding syllable, so that the high target is reached at the beginning of the syllable (pattern D´), after which the pitch starts to fall; in another variant (D´´), the tone is level high:
(19) a. Ai vor.'bit cu co.'pi.ii?
L+<H* L*H% Pattern C
L+<H* H+L*H% Pattern C´
b. Ai vor.'bit cu co.'pi.ii?
L* L+H* L% Pattern D L* H*+L L% Pattern D´
L* H* L% Pattern D´´
The results for this pair of examples is more balanced: a majority of informants chose the predicted pattern – C or C´ (i.e. LH) for (15) and D or some of its variants for (16) (i.e., L..HL). Three chose the LH pattern for (16), i.e., the unmarked pattern instead of verum (note that two of those also failed to choose the verum pattern for (14)). Two chose the verum pattern for (15), and one chose a peculiar pattern, in which the first accent is H*, like in the LH pattern, but the final configuration is L+H* L%, as in the case of verum focus (pattern E). Given the context provided in (15), it is likely that those who opted for the verum pattern imagined that the hearer was expected to have a talk with her children, although this was not explicitly stated in the instructions, like in (16).
Table 1: results of the experiment
Informant: Pattern for (13): (14): (15): (16):
1 A B E D
2 A B C D´´
3 A B D D
4 A A C D
5 A A C´ D´
6 A A´ C´ D
7 A A C D
8 A B C C
9 A B C D
10 A A C C
11 A B D D
12 A A C C
13 A A C D
Here are examples of the most characteristic patterns (figs. 5–6, for eg. (13)–
(14), and figs. 7–8, for eg. (15)–(16); for each contrasting pair, I used data from the same informant):
Figure 5: Pattern A
Figure 6: Pattern B
Figure 7: Pattern C
Figure 8: Pattern D
In order to support the claim that the rising accent on the verb in patterns A and C is a prenuclear accent, I tested a declarative version of the sentence in (13)- (14), in an out-of-the-blue context. All speakers produced a rising pitch on the verb (L+H* or L+<H*); in the end, most produced a falling pitch (H+L* L%), characteristic of unmarked declaratives (cf. Göbbel 2003, Jitcă et al. 2015); two produced a rise-fall pattern (H*+L L%), characteristic of exclamatives or declaratives with an exclamative flavor (Jitcă et al. 2015).
(20) [Context: you meet some acquaintances on the street by chance]
Văd că... aţi 'fost la cum.pă.ră.'turi.
see.1SG that have.2PL been at shoppings
L+(<)H* H+L* L%
‘I see.. you’ve been shopping.’
Figure 9
Concerning the LH pattern in questions (patterns A and C), in most of the previous literature it has been considered to reflect focus on the last phrase (Avram 1973, Dascălu 1979a,b, Dascălu-Jinga 2001, Ladd 2008). It is true that narrow focus on the last phrase results in this pattern, but in this study I have shown that the same pattern is found with truly neutral polar questions. As final narrow focus is usually realized in the same way as ‘broad focus’ (see Selkirk’s 1984 classical
‘focus projection rules’), except if special contrast is added, the fact that neutral polar questions and polar questions with narrow focus on the last constituent have the same prosodic realization is what we should expect given the general prosodic system of the language.
We should also note that the early focal accent on the verb, i.e. the L*..H(L) pattern, is not restricted to verum focus, it can also be due to the deaccenting of the postverbal material. This can be seen by comparing the intonation of the declarative and interrogative versions of the same clause, uttered in the context in (21); as the verb plăcea ‘like’ disallows thetic uses, being an i-level predicate, usually at least one of its arguments is contextually given, hence deaccented; in (21)a, the occasion of mentioning the subject’s enjoying Enescu’s music is provided by the occurrence in the context of a sample of this music; note that Enescu is deaccented, due to this relation with the context; the same context allows a question about the hearer’s feelings for Enescu’s music; we expect the same deaccenting, in (21)b. What we find is the same prosodic pattern as the one used with verum focus: L* on the verb and H* L% on the last constituent, Enescu:
(21) [Context: Enescu’s music is on the radio]
a. Îmi 'PLA.ce E.'nes.cu L+H* L*L%
‘I like Enescu’
b. Îţi 'PLA.ce E.'nes.cu?
L* L+H* L%
‘Do you like Enescu?’
3. DISCUSSION AND OPEN ISSUES
We have seen that the use of a nuclear stress on the main verb, realized like an early focus, does not represent the pattern of neutral polar questions, as claimed in previous studies, but is a genuine early focus pattern, signaling the high contextual accessibility or activation of the issue ‘whether p’, which corresponds to the semantic characterization of verum focus in questions (see Gutzmann and Castroviejo Miró (2011), Lohnstein (2016)). However, the fact that the L..HL pattern has been misinterpreted as neutral can show more than a mere failure of recognizing verum focus. It can be that the pragmatic conditions for using a verum focus intonation in questions are more relaxed in Romanian (and possibly in the
other eastern European languages which Grice et al. claim to behave similarly) than in English, or in other Romance languages. The context-givenness of the question can be a matter of degree. The condition in (9) states that the question q must already be among the questions under discussion (QUD), but the boundaries of this set are not clearly established. We have seen examples where the antecedent of q occurred in a previous linguistic exchange, separated in time from the dialogue at hand (see (11), (14), (16)) and even where there was no antecedent, but q was somehow expected to occur in the context (see (12)a). This is why I talked about
‘contextual accessibility/activation’ instead of ‘givenness’ at the beginning of this paragraph. It is possible that other languages may require a higher degree of activation for the use of verum focus. The existence of special ‘flavors’ of verum focus, which can differ from language to language and also across various constructions (see Giurgea and Mîrzea-Vasile forth. for another Romanian construction analyzable using verum focus, which expresses reassurance, threat or concession) favors accounts based on verum operators, whose meaning can be variously specified (see the proposals mentioned in section 1, in particular Gutzmann and Castroviejo Miró 2011, Gutzmann et al. 2017): if verum focus represented focus on a preexisting head, found crosslinguistically, we would expect it to manifest uniform pragmatic conditions across languages and constructions.
However, extensive comparative research is needed before drawing conclusions on this issue.
A further problematic case is the use of the L..HL pattern in interrogative requests. In the following example, no activation of ?q is required. The question is out-of-the-blue, it checks whether the hearer agrees with a proposal that the speaker makes using this sentence:
(22) [Context: during a walk, the speaker has the idea of visiting somebody who lives nearby]
Ai 'vrea să 'tre.cem pe la So.'ri.na?
would.2SG want SBJV pass.1PL by Sorina H* L* H* L%
‘Would you like to drop in on Sorina?’
The use of the neutral LH pattern here is appropriate if the question is a request of information about a preexisting intention of the hearer, in other words, if it is a genuine question rather than a proposal. The connection of the pattern in (22) with verum focus, or with early focus in general, is not at all clear (note also that the L* tone is not on the main verb, as expected for verum focus). This pattern might also explain (12)a – asking whether the seller has a certain stuff may be used as a way of requesting it.
The analysis of the L..HL pattern I proposed does not extend to (22). The contrast between L..HL representing verum focus and LH representing the neutral intonation holds for genuine requests of information. Finally, let us stress that the
verum L..HL pattern does not involve any expectation towards a positive or negative answer. The only difference with respect to the neutral pattern (LH) is the requirement of a certain degree of contextual activation of the question at hand.
REFERENCES
Avram, A., 1973, “Particularităţi ale intonaţiei interogative în graiul din Muscel”, Fonetică şi dialectologie 8, 43–64.
Barbu, A. M., 1999, “Complexul verbal”, Studii şi cercetări lingvistice, 50, 1, 39–84.
Bianchi, V., G. Bocci, S. Cruschina, 2015, “Focus fronting and its implicatures”, in: E. O. Aboh, J. C. Schaeffer, P. Sleeman (eds), Romance Languages and Linguistic Theory 2013. Selected papers from ’Going Romance’ Amsterdam 2013, Amsterdam/Philadelphia, John Benjamins, 1–19.
Bianchi, V., G. Bocci, S. Cruschina, 2016, “Focus fronting, unexpectedness, and evaluative implicatures”, Semantics and Pragmatics 9, 3, 1–54.
Bianchi, V., Cruschina, S., 2016, “The derivation and interpretation of polar questions with a fronted focus”, Lingua, 170, 47-68.
Boersma, P., D. Weenink. 2012. Praat – Doing phonetics by computer [Computer program].
http://www.praat.org/ [Retrieved on June 13, 2016].
Büring, D., 2006, “Intonation und Informationsstruktur”, in: H. Blühdorn, E. Breindl, U. H. Waßner (eds), Text-Verstehen. Grammatik und darüber hinaus, Berlin, New York, de Gruyter, 144–
163.
Büring, D., 2007, “Semantics, Intonation, and Information Structure”, in: G. Ramchand, C. Reiss (eds), The Oxford Handbook of Linguistic Interfaces, Oxford, Oxford University Press, 445–
473.
Büring, D., 2012, “Predicate integration – phrase structure or argument structure?”, in: I. Kucerova, A. Neeleman (eds), Contrasts and Positions in Information Structure, Cambridge, Cambridge University Press, 27–47.
Dascălu, L., 1975, “Observaţii asupra intonaţiei graiului din Maramureş”, Fonetică şi dialectologie 9, 77–91.
Dascălu, L., 1979a, “Teste de percepţie asupra intonaţiei în limba română”, Studii şi cercetări lingvistice 30, 2, 125–139.
Dascălu, L., 1979b, “On the intonation of questions in Romanian: the rising pattern”, Revue roumaine de linguistique 24, 1, 35-44.
Dascălu, L., 1986, “Asupra intonaţiei graiului din Bihor”, Studii şi cercetări lingvistice, 37, 3, 221–239.
Dascălu-Jinga, L., 2001, Melodia vorbirii în limba română, Bucureşti, Univers Enciclopedic.
Dik, S. C., M. E. Hoffman, J. R. de Jong, S. I. Djiang, H. Stroomer, L. de Vries, 1981, “On the typology of Focus Phenomena”, in: T. Hoekstra, H. van der Hulst, M. Moortgat (eds), Perspectives on Functional Grammar, Dordrecht, Foris Publications, 41–74.
Dobrovie-Sorin, C., 1994, The Syntax of Romanian, Berlin, Mouton de Gruyter.
Dobrovie-Sorin, C., Galves, C., 2000, “Proclisis, Enclisis and Head-to-head Merge”, Bucharest Working Papers in Linguistics 2, 1, 35–51.
Frota, S., Prieto, P. (eds.), 2015, Intonation in Romance, Oxford, Oxford University Press.
Giurgea, I., 2011, “The Romanian verbal cluster and the theory of head movement”, in: J. Herschen- sohn (ed.), Romance Linguistics 2010. Selected papers from the 40th Linguistic Symposium on Romance Languages (LSRL), Amsterdam, Philadelphia, John Benjamins, 271–286.
Giurgea, I., 2016, “On the interpretation of Focus Fronting in Romanian”, Bucharest Working Papers in Linguistics, in press.
Giurgea, I., C. Mîrzea Vasile, forth., “Syntactic effects of verum focus in Romanian”, Revue roumaine de linguistique.
Giurgea, I., E. Remberger, 2012, “Verum Focus and Polar Questions”, Bucharest Working Papers in Linguistics, 14, 2, 21–40.
Giurgea, I., E. Remberger, 2014, “Polarity Fronting in Romanian and Sardinian”, in: M.-H. Côté, E. Mathieu (eds), Variation within and across Romance Languages. Selected papers from the 41st Linguistic Symposium on Romance Languages (LSRL), Ottawa, 5–7 May 2011, Amsterdam, Philadelphia, John Benjamins, 173–197.
Göbbel, E., 2003, Syntactic and Focus-structural Aspects of Triadic Constructions, Frankfurt, Peter Lang.
Grice, M., D. R. Ladd, A. Arvaniti, 2000, “On the place of phrase accents in intonational phonology”, Phonology, 17, 143–185.
Gussenhoven, C., 1984, On the Grammar and Semantics of Sentence Accents, Berlin, New York, Walter de Gruyter.
Gutzmann, D., E. Castroviejo Miró, 2011, “The Dimensions of Verum”, in: O. Bonami, P. Cabredo Hofherr (eds.), Empirical Issues in Syntax and Semantics 8, http://www.cssp.cnrs.fr/eiss8, 143–165.
Gutzmann, D., K. Hartmann, L. Matthewson, 2017, “Verum focus is verum, not focus. Cross- linguistic evidence”, ms., available at https://www.danielgutzmann.com/work/verum-focus-is- verum-not-focus.
Guţu-Romalo, V., 1962, “Forme verbale compuse”, Studii şi cercetări lingvistice, 13, 2, 191–199.
Halliday, M. A. K., 1967, Intonation and Grammar in British English, The Hague, Mouton.
Höhle, T. 1988, “Vorwort und Nachwort zu Verumfokus”, Sprache und Pragmatik 5, 1–7.
Höhle, T. 1992, “Über Verumfokus im Deutschen”, in: J. Jacobs, (ed.), Informationsstruktur und Grammatik, Opladen, Westdeutscher Verlag, 112–141.
Jitcă, D., V. Apopei, O. Păduraru, S. Maruşca, 2015, “Transcription of Romanian intonation”, in:
S. Frota, P. Prieto (eds.), Intonation in Romance, Oxford, Oxford University Press, 284-316.
Ladd, R., 2008, Intonational Phonology (2nd edition), Cambridge, Cambridge University Press.
Lohnstein, H., 2012, “Verumfokus—Satzmodus—Wahrheit”, in: H. Lohnstein, H. Blühdorn (eds.), Wahrheit—Fokus—Negation, Hamburg, Buske Verlag, 31–67.
Lohnstein, H., 2016, “Verum focus”, in: C. Féry, S. Ishihara (eds), The Oxford Handbook of Information Structure, Oxford, Oxford University Press, 290–313.
Pierrehumbert, J., 1980, The phonetics and phonology of English intonation, PhD dissertation, MIT.
Pierrehumbert, J., J. Hirschberg, 1990, “The meaning of intonational contours in the interpretation of discourse, in: P. Cohen, J. Morgan, M. Pollack (eds), Intentions in Communication, Cambridge, Mass., MIT Press, 271–311.
Repp, S., 2013, “Common ground management: Modal particles, illocutionary negation and verum”, in: D. Gutzmann, H. Gärtner (eds), Beyond Expressives. Explorations in Use-Conditional Meaning, Leiden, Brill, 231–274.
Rochemont, M., 2013, “Discourse new, F-marking, and normal stress”, Lingua, 136, 38–62.
Rochemont, M., 2016, “Givenness”, in: C. Féry, S. Ishihara (eds), The Oxford Handbook of Information Structure, Oxford, Oxford University Press, 41–63.
Romero, M., Ch. Han, 2004, “On negative y/n-questions”, Linguistics and Philosophy, 27, 609–658.
Romero, M., 2005, “Two approaches to biased yes/no questions”, in: J. Alderete, C. Han, A. Kochetov (eds), Proceedings of the 24th West Coast Conference on Formal Linguistics (WCCFL), Somerville, MA, Cascadilla Proceedings Project, 352–360.
Selkirk, E., 1984, Phonology and syntax: the relation between sound and structure, Cambridge, MA, MIT Press.
Zimmermann, M., 2016, “Predicate focus”, in C. Féry, S. Ishihara (eds), The Oxford Handbook of Information Structure, Oxford, Oxford University Press, 314–335.
