Abstract
In this Chapter we will analyze the phonological structures of Focus and Topic within the Prosodic Phonology framework. Taking the application of phonological rules as diagnostics, our aim is to identify the minimal prosodic domains containing Focus and Topic constituents. By ‘minimal’ we mean the lowest prosodic domain in the prosodic hierarchy which exhaustively contains a Focus or a Topic, independently of their length. As mentioned in Chapter 1, this analysis — besides having intrinsic interest for phonological issues — will provide the basis for empirically grounded hypotheses about the syntactic structures of Focus and Topic (to be tested in Chapter 3).
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Notes
For the distinction between lexical and functional categories and its role in characterizing the syntactic properties of the sentence, see Chomsky 1986a, Abney 1987, Pollock 1989, Grimshaw 1991, among others.
Though we basically agree with Nespor and Vogel’s (1986) definitions of the domains of application of phonological rules (but see note 4), we sometimes move away from the SPE formalisms (Chomsky and Halle 1968) the authors adopted to represent them; instead, we prefer to use Autosegmental formulations (Goldsmith 1990).
Most of the work on RS refers to the Florentine variety (cf. Marotta 1986, Trumper etal. 1991). The variety of RS spoken in Rome is basically similar to the Florentine one (cf. Chierchia 1986), differing only in its monosyllables, and a small group of bisyllable paroxitones (such as come “how”, dove “where”). The choice of Florence, Rome and Milan as sample-towns for the analysis is in line with the selection used in the LIP database (De Mauro, ed. 1993), whose data will be used in our syntactic analysis.
The requirement of [+son, -nasi features in (7) excludes the possibility of RS in sequences like [pn], so that no lengthening of [p] can be found in venderti pneumatici (“he will sell tires”), while RS is present in venderti [p:]lastica (“he will sell plastic”) or venders [p]rugne (“he will sell plums”).
However, the behavior of the Italian `[s] impura’, involving the clusters [sr] and [sl], complicates this formulation. One condition should be added in the spirit of the Licensing Approach of the Autosegmental Theory (cf. Goldsmith 1990), namely that “aa2 presents a licensed onset”. This would exclude the extrametric [s] impura from RS application (as is known, illegitimate elements are invisible to phonological rules; on this point cf. also Marotta 1995). A more precise formulation of RS in Autosegmental terms will be the subject of future research.
Finally, the DTE abbreviation stands for `Designated Terminal Element’; it defines the strong terminal node which dominates the syllable bearing the main stress of the word.
This rule is often called `Stress Retraction’ in the Italian literature. Since its application does not trigger any prominence shift at all, we prefer to use the definition of Rhythm Rule, as used by Hayes (1989) and Gussenhoven (1991), among others. For an analysis of RR in other languages, cf. Liberman and Prince (1997), Hayes (1984), Hayes and Puppel (1984), Kager and Visch (1985).
Indeed, the two rules have almost identical contexts of application and the analysis of data shows that speakers of central and southern varieties apply both rules in the relevant contexts. The fact that the RR has more phonetic relevance in northern Italian varieties can be plausibly attributed to the lack of RS in these varieties, which somehow `obscures’ RR in central and southern Italian. However, for the sake of clarity, the application of RR and RS will be illustrated in separate examples in the analysis which follows.
In the examples presented here, Italian orthography is not followed as far as accent marks are concerned. Word-final stress does not show the acute-grave accent mark alternation but is always marked as acute (following general standards in the phonological literature). Word-internal stress is also marked, whenever relevant.
We have not graphically distinguished possible spirantized forms of the phonemes [h, 0, cp] realized by our speakers (a continuum which approaches complete vanishing in the case of [h]). It is important to remember that we always refer to different degrees of spirantization in the analysis. Moreover, GT incidence is not the same for all occlusive consonants: there is an asymmetric relationship between [h] and [0, (p]. This asymmetry led scholars of diachronic linguistics to hypothesize that spirantization of the velar occlusive occurred before or extended more rapidly than the others (cf. Giannelli and Savoia 1978, 1979–80).
The Discourse-Configurationality parameter distinguishes languages which realize some discourse categories in an obligatory and systematic way from those which do not. Languages can thus be defined as `Focus-prominent’ (e.g., Hungarian), `Topic-prominent’ (e.g., Japanese) or `Subject-prominent’ (e.g., English); for details, cf. Kiss, ed. (1995a).
Branching constituents dominate more than one constituent of the immediately inferior prosodic level, whereas non-branching constituents immediately dominate only one.
It is the specific constituent Kanerva (1990) assumes to contain a Focus after the restructuring process. It can fundamentally be considered the same entity as a Phonological Phrase (cf. Truckenbrodt 1999).
The phonological rules used as diagnostic are typically met in the varieties spoken in the three mentioned towns. Specifically, Rome for RS and (partially) for IS, Florence for RS, IS and GT. RR was particularly tested in the Milan variety, but it was also examined for all other speakers (see discussion in Section 2.1.1 and note 6).
The reader will note that no examples with [+F] Verbs are offered in this section. This is deliberate since the prosodic behaviour of [+F] Verbs is an interesting point of discussion and will be treated in detail in Section 2.4.4.
The phonological nature of Focus Restructuring can be interpreted in the light of a general wellformedness condition operating on prosodic constituents whereas its direction follows the principles of Metrical Phonology (this will be discussed in Section 2.5 below).
This can be plausibly explained by two factors: a) the default main prominence assigned to the rightmost element within the sentential I and b) a well-formedness condition operating on prosodic constituents (see discussion in 2.5.1).
For a syntactic analysis of postverbal Subjects and relevant references, see Chapter 3, Section 3.1.3.3.
The peculiar intonation associated with contrastive Focus must thus be considered a later effect in the PF component; this will not be investigated here as it is not relevant to the main argument.
It is important not to confuse general (13-restructuring with Focus Restructuring. The former is never obligatory and, more important, never disrupts the prosodic phrasing determined by general algorithms (i.e., by the Mapping Rules). Focus Restructuring, on the other hand, is obligatory and has the overriding power to re-organize the prosodic phrasing.
This unusual behaviour of the Verb and its role in Focus constructions will be analyzed in Chapter 3.
We deliberately keep syntactic considerations very vague so as not to disrupt the issues of prosodic analysis.
Extrasyllabicity is particularly invoked to account for the behavior of marginal consonants in escaping syllabification and of final vowels in escaping tone spread. Extrametricality was also proposed to explain why peripheral syllables can skip metrical parsing rules (cf. Kenstowicz 1994, 274, 567). Given the existence of this phenomenon in our data, what we propose here is to extend its application to postlexical prosodic levels.
It has been proposed that linguistic material may remain unparsed until postlexical Phonology (cf. Kenstowicz 1994, 285ff.). Given the present analysis, we propose, in agreement with Rice (1990), that extraprosodicity is allowed to tum off at levels other than the word level. This point, however, is not relevant here.
The repetition of the given constituent in the answer is, in fact, rather unnatural, but it is necessary here to test the application of phonological rules on the right side of the new Focus.
The crucial role of a left-to-right vs. right-to-left parameter in the application of phonological rules is widely attested. Consider, for example, its role in the construction of metrical grids accounting for a myriad of stress systems (cf. Prince 1983, Halle and Vergnaud 1987, Hayes 1991 and, for general discussion, Kenstowicz 1994).
But see Section 2.4.4 for the case of non-branching Topics which are left-adjacent to (syntactic) contrastive Focus constituents.
It goes without saying that I-domain rules are present within Topic constituents, but their phonetic representation is omitted since only limit domain phenomena are relevant to the present discussion.
Although Topicalization is available in English (though with severe restrictions, cf. Section 3.2.6.1), here and in the rest of the book we have preferred not to use Topicalized structures in translations because, given the constraints ruling this constructions in English, reproducing the Italian order using expressions like “as forchwr(133)” and so on produces awkward and unnatural results (especially in multiple Topic constructions). So the location of Topic constituents rests on glosses only.
The defective nature of non-branching nodes is generally recognized in the literature (cf. Goldsmith 1990 for important discussions). We will retum to this point in Section 2,5.
Chierchia (1986) argues for a cross-linguistic `Must Branch Condition’ for which different prosodic constituents are parametrically selected as obligatorily branching across languages. Our analysis seems to proceed in this direction.
As noted by a reviewer of this series, this conclusion seems to be contradicted by Bengali which, though a left-recursive language, locates the main prominence on the rightmost edge of the sentential I. For this reason he/she concludes that principles governing the metrical trees below the word level cannot be extended to higher constituents. We do not agree with this objection. In fact, metrical principles seem to govern prosodic levels up to the (t level in Bengali as well, and, if a line must be drawn to exclude a prosodic level from general metrical principles, this should single out only the I-level. As stated in Hayes and Lahiri’s (1991) paper, “within the P-phrase [i.e., the N], the leftmost nonclitic word is the strongest” (p. 55). This is clear evidence that from a phonological point of view Bengali is a left-headed language — at least up to the <D-level. Consequently, Bengali Focus Restructuring, which is a 4>-level phenomenon, is consistent with our analysis since it would proceed in a direction which is opposite to syntactic recursion in Bengali (i.e., rightwards). We do not have original Bengali material to work on but, as far as we can see from Hayes and Lahiri examples, a [+F] does have a boundary on its left side (the recursive side) which is not present in broad Focus sentences and which blocks the application of rules which would be otherwise present (pp. 82–83). This is consistent with our analysis. At any rate, the presence of the main prominence on the rightmost t) within the sentence seems to be the only `deviation’ to metrical geometry and the authors themselves tend to ascribe it to a “true phonological rule” (p.93), a sort of natural phonological condition which requires the main stress to be at the end of an intonative group. This is also consistent with our Prosodic Null Theory (121), that is to say main prominence is assigned to the rightmost legitimate constituent within the sentential I, independently of the Head Parameter.
We hinted at the split in the CP node in Chapter 1 (Section 1.1). Nonetheless, we will discuss this point in more detail in Chapter 3.
Cinque (1993) acknowledges these problems and concludes that two grammars must be distinguished: a sentence grammar and a discourse grammar, the latter changing the output of the former in marked Focus assignment. Radically proceeding in this direction, Reinhart (1995) claims that the computational system derives only a “set of possible foci” and marked Focus operations are governed by discourse appropriateness. Reinahrt defines Focus as a “relation between a structure <PF, LF> and a context” (p. 75). But, as Reinhart herself observes, this seems to be too permissive a system and the problem is how and where to restrict it.
Our amendment on trace visibility for the purposes of main prominence assignment is in line with Selkirk’s (1993) Focus Projection Rule. The author provides arguments in its favor, showing that it can explain a number of classic puzzles about main prominence. She states that “F-marking of the antecedent of a trace left by NP- or wh-movement licenses the F-marking of the trace”. Trace visibility was also supported elsewhere in the literature, for different reasons (cf., among others, Rizzi 1979, Chomsky 1981).
Our proposal can also provide a prosodic interpretation for the syntax-based theory of Focus Projection (cf. Selkirk 1993, Zubizarreta 1994, among others). This states that F-marking can project from the [+F] word to all the arguments in the sentence. Prosodic Null Theory defines the domain of this projection, i.e., the sentential I whose right boundary is marked by the strong O.
Hausa is an Afroasiatic language of the Chadic group. Data presented in this section is taken from Vogel and Kenesi (1990) and Zec and Inkelas (1990).
The position of the proper Noun Audu in (122) can be interpreted in the light of Longobardi’s (1994) ‘N-to-D movement’ of non-quantified nominal constituents (see also Chapter 3, note 65).
Since syntactic Focus is the unmarked option in Hausa, the fact that in situ Focus — as the marked option — must receive emphatic intonation is consistent with general cross-linguistic data. Italian, for example, is a mirror case: in situ Focus (the unmarked option) does not require any special melody, while syntactic Focus must be marked by a specific emphatic/contrastive intonation.
Sentence (130) is thus ruled out as a case of multiple Focus.
Chichewa is a Bantu language spoken in East Central Africa, particularly in Malawi, Mozambique, Zambia and Zimbabwe. Examples reported in this section are taken from Kanerva (1990). The author does not provide glosses; consequently they are missing here as well.
However, to consider all scrambling phenomena as `prosodic strategies’ encounters several problems which we will discuss in Chapter 3. For now we will only consider the prosodic consequences of this kind of analysis.
Data presented in this section are taken from Zec and Inkelas (1990).
Serbo-Croatian, like a number of languages, locates clitics in second position (what is known as `Wackemagel law’). `First’ positions can be occupied by lexical words (Nouns, Verbs, Adjectives) or by maximal categories. So, even though taj cbvek and Petar — in (144) and (145) respectively — may plausibly be considered to be the Subjects of the relevant sentences, their Topic role derives from the presence of the clitic (the auxiliary je). The clitic identifies the Verb as the first constituent of both sentences thereby triggering the extrasentential position of what precedes it.
This analysis thus also supports the need for the C level in the prosodic hierarchy (see Chapter 1, note 9).
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2000 Springer Science+Business Media Dordrecht
About this chapter
Cite this chapter
Frascarelli, M. (2000). Prosodic Analysis. In: The Syntax-Phonology Interface in Focus and Topic Constructions in Italian. Studies in Natural Language and Linguistic Theory, vol 50. Springer, Dordrecht. https://doi.org/10.1007/978-94-015-9500-1_2
Download citation
DOI: https://doi.org/10.1007/978-94-015-9500-1_2
Publisher Name: Springer, Dordrecht
Print ISBN: 978-90-481-5426-5
Online ISBN: 978-94-015-9500-1
eBook Packages: Springer Book Archive