Show Summary Details

Page of

Printed from Oxford Research Encyclopedias, Linguistics. Under the terms of the licence agreement, an individual user may print out a single article for personal use (for details see Privacy Policy and Legal Notice).

date: 01 October 2022

Hanging Topics and Frames in the Romance Languages: Syntax, Discourse, Diachronyfree

Hanging Topics and Frames in the Romance Languages: Syntax, Discourse, Diachronyfree

  • Elisabeth StarkElisabeth StarkInstitute for Romance Studies, University of Zurich


Hanging topics and frames are optional, adjunct-like utterance-initial elements without any syntactic function inside the clause they precede. Both terms are frequently used in an ambiguous way in the specialized literature, in a way that often confounds syntactic and functional properties. However, hanging topics and frames can be kept apart. Hanging topics, on the one hand, are defined as utterance-initial syntactically and often prosodically independent constituents that denote the topic referent, that is, a discourse referent, an information element comparable to a file card under which the comment, that is, the related information provided in the following sentence, has to be stored (“aboutness”). Hanging topics are thus one type of topic-marking construction, alongside dislocations, which are, however, syntactically more dependent on the clause they precede or follow. Frames, on the other hand, are syntactically even more independent than hanging topics: they are not coreferential to any element of the accompanying sentence, and they cannot be integrated in the following sentence without changing their scope behavior. Additionally, their function is different: Rather than denoting the topic of the following utterance (there is, however, a subtype that does so and is thus to be classified between hanging topics and frames), they denote or delimitate the interpretational frame (‘domain indication’) for the following utterance. Both constructions show a rather neat correlation between the discourse-pragmatic status of their referents as given or new and the prosodic and categorial marking: The newer the discourse referent, the more prominent its intonational profile and the more likely the presence of thematic markers (like Fr. quant à, ‘as for’). In a diachronic perspective, hanging topics and frames, constituting universally available initial elements of utterances, whose use is mainly coherence driven, do not show considerable changes from Latin to the Romance languages in terms of their syntax or morphophonology. What has basically changed, to a different extent in different Romance languages, is their variationist markedness (from colloquial to standard registers in some cases). In fact, hanging topics and frames have always been available in Romance as well as in Latin, where they are known as instances of nominativus and less frequently also as (adverbial) accusativus pendens.


  • Pragmatics
  • Syntax

1. Introduction

This chapter deals with two types of constituents “outside” the clause in Romance, covering two different discursive functions, that is, hanging topics and frames. Both terms seem to have originated in the 1970s, in functional and formal approaches to sentence-initial syntactic phenomena closely linked to information structure (cf. e.g., Chafe, 1976; Cinque, 1977) and marked with respect to neutral constituent orders, in Romance SVO. Hanging topics are also discussed, especially in earlier literature, under the term of nominativus pendens. Take the French examples (1)(3):




The noun phrases in bold have received different names in the specialized literature: the one in (1) is known as (clitic) left-dislocation (ClLD); the one in (2) is known as hanging topic (HT), both at least since Cinque (1977, 1979); and the one in (3), albeit not consistently, is known as (absolute) frame (AF; cf. e.g., Stark, 1997).

In most studies on the topic, ClLD and HT are considered to be neatly distinguishable on the basis of several syntactic properties that can be summed up under the notion of “connectivity” (cf. Cinque, 1997 and Section 2.1), while they both serve as topic-marking constructions.

The information-structural notion of topic is linked to aboutness (cf. for the first systematic definition and analysis Gundel, 1975, 1977; see also article on “Topicalization in the Romance Languages” in this encyclopedia) and will be defined for this article as follows:

The topic constituent identifies the entity or set of entities under which the information expressed in the comment constituent should be stored in the C[ommon] G[round] content.

(Krifka, 2007, p. 41)

This definition continues the original library metaphor used by Reinhart (1981) to capture the prototypical function of topic expressions in discourse (see also e.g., Vallduví, 1992) and seems to privilege the “addressation function” of topics argued for by Jacobs (2001, p. 650), one of the four semantic attributes of which the notion of topic can be composed, according to him (alongside informational separation, predication and frame-setting; we will return to the latter aspect in Section 2.2).

The Common Ground (CG), in Krifka’s (2007) terminology, is the information shared by speaker and hearer in a linguistic interaction, and topics serve to indicate under which “file” (cf. Heim, 1982, Kamp & Reyle, 1993, for similar modeling of discourse semantics) incoming information should be registered. In example (1), the CG would be les malotrus (‘louts’), about whom it is stated in the corresponding comment constituent that they are not (generally) invited; in (2), the topic referent is a person called Claas (= topic constituent), whose shoes the speaker has taken a picture of. Topic-marking is one of the key operations of C[ommon] G[round] management (in Krifka’s [2007] terms) or of “information packaging” in an early approach to information structure by Wallace Chafe (1976). The “activation status” (cf. e.g., Ariel, 1991; Baumann & Riester, 2012; Givón, 1992; Prince, 1981, 1992) of a topic referent can differ; it can be new or unfamiliar to the hearer or easily identifiable, because it has been recently mentioned, is present in the extralinguistic context, or part of the shared or general knowledge of speaker and hearer. There are thus different shades of “activation” or “givenness”, usually represented in scales. For what follows, a very rough distinction between “given” and “new” topic referents will suffice, and more subtle descriptions will only be resorted to when the linguistic data discussed require them.

The initial constituent les autres pays (‘other countries’) in example (3) differs from those in (1) and (2) in that a description of its function by ‘aboutness’ is less appropriate. The following sentence c’est un bout de papier plastifié (‘it is a laminated piece of paper’) does convey information about a (topic) referent, but not about les autres pays, ‘other countries’—it qualifies their respective identity cards, which are not mentioned, but are inferrable by the implicit contrast with the Swiss identity card at the beginning of the preceding sentence. In order to capture the discourse function of this type of constituent, Chafe (1976, p. 50) called them “Chinese-style topics” and stated that they set “a spatial, temporal, or individual framework within which the main predication holds.”1 Krifka (2007) calls these expressions “frame setters” and points to their adverbial nature in a broad sense; the kind of information that can appropriately be given in the following utterance depends on the frame set, taking up in almost identical terms the description of frame-setting proposed by Jacobs (2001, p. 656). Aboutness in the sense of “storage adressation” is thus not at issue in frames or frame-setting constituents (see also Jacobs, 2001, p. 654, example 23, and p. 658). These are sometimes marked by explicit discourse particles such as Ger. was . . . betrifft, Eng. as for, Fr. quant à, It. per quanto riguarda, Pt.: no que diz respeito a, Cat. pel que fa a, quant a, and Ro. în ceea ce privește, cu privire la, but the presence of these elements does not constitute a sufficient criterion to distinguish frame setters from topic constituents (see Section 4; see also Dolci, 1986; Escobar, 1995; Rivero, 1980; Silva Corvalán, 1984; or Zubizarreta, 1998, p. 187 for an analysis of as for constructions as HTs; see e.g., Larsson, 1979 or Villalba, 2000, 2009 for their analysis as a topic construction on their own). Most importantly, this notion of frame, referring to a syntactically isolated constituent at the beginning of sentences and indicating the domain in which the following predication is valid, is not to be confounded with the semantico-conceptual notion of cognitive frame or knowledge frame in the tradition of Schank and Abelson (1977) (cf. Bednarek, 2005; Minsky, 1977; Tannen, 1979; see also Blackwell, 2017).

In what follows, it will be shown that the notions of HT and frame are both polysemous, as they may denote a specific syntactic constituent type (formal criteria, see Table 1) or two different functional types of entities in an utterance (see Table 2), with formal and functional criteria that do not unambiguously match. This leads to the following cross-classification (see Table 3), which will guide the overview of some key Romance linguistic studies on the issue in order to clarify terminological pitfalls, but also to understand and correctly evaluate seemingly contradictory findings and analyses:

Table 1. Formal Criteria to Distinguish Hanging Topics and Frames (First Approximation)

Formal properties

Hanging Topic


Initial position



Nominal constituent



Coreferential element possible (except clitics)



Table 2. Functional Criteria to Distinguish Hanging Topics and Frames

Functional properties

Hanging Topic





“Domain indication”



If only the two distinguishing properties of the (potential) presence of a coreferential element and “aboutness” (cf. Velghe & Lahousse, 2015, p. 427, p. 439 for French) are taken into account, the following picture arises:

Table 3. The Four Main Types of Hanging Topics and Frames



“Domain indication”

Coreferential element possible (except clitics)

Hanging Topic


Coreferential element impossible

“frame-like HT” French TDc’est*


* Note: See Section 2.2.

The cell highlighted in gray in Table 3 underlines the observation expressed very clearly by Jacobs (2001, p. 641): Left-peripheral topic constituents may be integrated into (or connected to) the following sentence by the presence of a coreferential element via co-indexing, or nonintegrated by the absence of such an element. In fact, there exists one often-overlooked type of initial topic-marking construction that corresponds syntactically to the definition of a frame, but that functionally marks discourse referents “about which” the main predication holds, such as votre sucre (‘your sugar’) in example (4):


The notion of the HT is thus most prone to terminological confusion (see also Prévost, 2003, p. 52), which will be avoided in the following sections by separating formal (Section 2) and functional (section 3) definitions of HTs and frames, while maintaining these two well-established terms.

After this introduction, I will discuss formal (Section 2) and functional (Section 3) properties of the two structures (HTs and frames). Further relevant aspects such as diachrony, markedness, and thematic markers are discussed in Section 4. Section 5 provides a short conclusion.

2. Formal Description

2.1 Hanging Topics (HTs)

As shown in the introduction, hanging topics (HTs) will be understood here as utterance-initial constituents that denote the topic referent; that is, a discourse referent, an information element comparable to a file card under which the comment, that is, the related information provided in the following sentence, has to be stored (“aboutness”). HTs are thus to be classified under topic-marking constructions. At least since Cinque (e.g., 1977, 1979, 1997), and most prominently for Italian, mainly two sentence-initial topic-constructions have been distinguished on the basis of their connectedness or syntactic dependency with respect to the following sentence: hanging topic (HT)3 and clitic left-dislocation (ClLD; see e.g., Cinque, 1977 for Italian, Villalba, 2000, 2009 for Catalan; see also López, 2016, pp. 403–404 or Frascarelli, 2017, for a Romance overview). The pragmatic functions associated with these two constructions may differ from language to language (cf. Cinque, 1997, p. 94), but the syntactic criteria set up by Cinque provide robust syntactic evidence to clearly distinguish a connected (potentially moved, see Alexiadou, 2006, Section 4, for further discussion) constituent (ClLD) from a syntactically independent one (HT). Cinque (1997, p. 96) provided a comprehensive overview of the main differences between HT and ClLD constructions, all valid in Romance, where, as seen in Table 4, properties b and c are most relevant for the syntactic position of HT vs. ClLD constituents, and properties e and f for the distinction between moved, or having two potential positions in the sentence, and merged, or having only one position in the sentence; or, in other terms, being syntactically dependent vs. independent constituents:

Table 4. HTLD vs. ClLD constructions

(Some) Properties of HTLD

(Some) Properties of ClLD

(a) The left-hand phrase can only be an NP.

(a) The left-hand phrase can be an NP, PP, AP, or S (essentially any X maximal, in the sense of X’ theory).

(b) There may be at most one left-hand phrase.

(b) There is no (theoretical) limit to the number of left-hand phrases.

(c) The left-hand phrase occurs typically to the left of a “root” S.

(c) The left-hand phrase can occur to the left of “root” and “non-root” Ss.

(d) The “resumptive element” can be a “pronominal” noun (or epithet, like that poor guy) or an ordinary pronoun, either stressed or clitic.

(d) The “resumptive element” can only be a clitic pronoun.

(e) There is no connectedness between the left-hand phrase and the resumptive element (in terms of case-matching, etc.).

(e) There is obligatory connectedness between the left-hand phrase and the resumptive element (in terms of case-matching, etc.).

(f) The relation between the left-hand phrase and the resumptive element is not sensitive to island constraints.

(f) The relation between the left-hand phrase and the resumptive element is sensitive to island constraints.

Adapted from Cinque (1997, p. 96).

Cinque (1997, p. 97) illustrates property b with the following example:


The two initial constituents are not taken up by a resumptive clitic in the following sentence, so they obey Cinque’s HT criteria for resumption (property d). According to him, at least in Italian, there cannot be more than one HT constituent. Furthermore, HT constituents can only be found in front of root sentences or subordinate sentences with a proper illocutionary force (cf. e.g., Benincà, 2001; Bianchi & Frascarelli, 2010; Villalba, 2000, 2009, and many others):


Bouzouita (2014, pp. 28–29)4 assumed both properties also for Spanish HT vs. ClLD constructions, but De Cat (2007a, 2007b) contested them for French (see example [7] for two HT constituents in front of a sentence, and [8] for an HT constituent preceding an embedded sentence with presupposed information):



Properties e and f concern a whole bundle of morphosyntactic and semantic features which distinguish HT from ClLD constructions (according to López, 2016, p. 404, these two properties are the central ones to support this distinction), especially reconstruction effects and extractability of constituents out of certain subordinate contexts (such as adverbial adjuncts/clauses or relative clauses inside complex noun phrases, or NPs), known as strong syntactic islands since Ross (1967). For the latter, Frascarelli (2017, p. 482, footnote 14) provides the following examples:



The coreferential elements, which are not obligatory but acceptable in HT constructions, are found in adverbial adjuncts (example 9b) and adnominal relative clauses (example 10b) in Italian and French without leading to ungrammaticality. This does not hold for ClLD constructions (examples 9a, 10a), recognizable by overt case-marking or case-matching in Cinque’s (1997) terms (see above, property e in his table; see Villalba, 2000 for similar observations in Brazilian Portuguese and Catalan). Bouzouita (2014, p. 30), taking up this distinction and following work by Casielles-Suárez (2004); Escobar (1995), and Pablos (2006) on Spanish, gives the following example illustrating island-sensitivity (the island being a relative clause here) of ClLD as opposed to HT in Spanish:


As for reconstruction effects, a cover term for the semantic interpretation of constituents not at their actual surface position, but at the position with which they are associated (= where they would be located in unmarked word order, cf. Chomsky, 1977), López (2009) showed that ClLD (here: with the differential object marker a), but not HT constructions (= without case-marking) in Spanish show restrictions on coreferentiality (“Principle C-effects” in generative terms):


In (12a), el muy tonto cannot be read as an epithet of al árbitro, since this preposed element is interpreted as being in its postverbal object position, where as a referential expression it cannot be interpreted as coreferential with the subject of the clause. In (12b), el muy tonto can be interpreted as an epithet of el árbitro, which is taken as evidence for its basic initial position in the left periphery (= not only superficially, but also syntactically, el árbitro precedes el muy tonto, which licences corefentiality). Interestingly, French does seem to exclude reconstruction effects for HT as well as for apparent ClLD constructions (cf. De Cat, 2007a, 2007b):


In fact, De Cat (2007a, 2007b) provided evidence for the inexistence of ClLD constructions in contemporary French for Cinque’s properties b, c, e, and f, the properties not shared as a subset by ClLD and HT constructions and thus those that make it possible to distinguish any case, which López (2016, pp. 408–409) interpreted as a diachronic loss of ClLD in French. De Cat (2007b, pp. 489–490)5 also discussed the possibility of having HT constituents after a complete sentence in French (“H-type right dislocations”, see also López, 2016, p. 410), but the bulk of her examples contain clitic resumptive pronouns, a property shared with ClLD/clitic right-dislocation (ClRD), and she also points out differences, for example with respect to the necessity of case-marking the topic constituent (De Cat, 2007b, p. 515), so this question is left open for further discussion.

To sum up, HT constituents can be defined as follows from a formal perspective:


As for the prosodic properties of HTs, it has been found that in a variety of Romance languages (for French: Delais-Roussarie & Feldhausen, 2014; Delais-Roussarie et al., 2004; Rossi, 1999; for Italian: Frascarelli, 2000; Feldhausen, 2014a; Gili Fivela, 1999; for Spanish: Feldhausen, 2014b, 2016a, 2016b; Feldhausen & Lausecker, 2018; for Portuguese: Frota, 2000; for Catalan: Feldhausen, 2010), dislocated topic constituents as well as HTs are usually separated from the rest of the utterance by an intonation phrase (IP) boundary (the highest prosodic constituent over phonological phrases, PhP, and prosodic words, following Selkirk (1984), cf. López, 2016, pp. 410–412, Nespor and Vogel (1986); for a more nuanced view on the facts see Feldhausen (2016a, 2016b), and Delais-Roussarie and Feldhausen (2014), who deal with the question of how “strong” the prosodic boundary can be). Interestingly, at least for French, not so many differences in prosodic phrasing are found between ClLD constituents and HTs (De Cat, 2007b, p. 505; see also Delais-Roussarie et al., 2004 for French and Feldhausen, 2016a for Spanish).

2.2 Frames

Frame constituents are syntactically even more isolated than HT constituents. While they share the properties of being usually sentence-initial and not connected to the following sentence by case-marking, agreement, extraction restrictions (island effects) or the absence of reconstruction effects, they show additional features of nonintegration: They are not coreferential to any element of the accompanying sentence, and they cannot be integrated in the following sentence without changing their scope behavior (cf. Gergely Horváth, 2018, pp. 87–91, for French):




Tre mesi più tardi (‘three months later’), a frame nominal, is ambiguous in (15), as it can denote the moment of decision as well as the wedding date, a reading excluded in (16), in the prototypical initial frame position, and in (17), in a parenthetical, maybe orphan position (cf. Haegeman, 2009).

Prosodically, frame constituents are more independent from the following utterance than syntactic HTs (and ClLD constituents); they have their own intonational contour (IC), but very often also a neat prosodic break following them is perceptible, in many cases a pause (cf. Stark, 1997, chapter 3 for empirical details observed in her corpus of contemporary spoken French).

If one applies these formal characteristics to a corpus analysis, at least two different functional types of frames will be identified. The first one (see Jacobs, 2001’s ‘frame-setting’ function or Frascarelli, 2017’s “L(imiting)-topic”) is illustrated in example (16) by initial adverbial nominals, not especially marked stylistically, at least not for temporal adverbials, which very often consist of simple noun phrases. Lexical changes can, however, induce a nonstandard evaluation of the respective constructions. Compare the following French example:


If retour is replaced by fois (‘time’), a fully acceptable initial temporal adverbial setting the temporal frame for the following proposition is obtained (la première fois, je devais avoir deux ans). The same “standardizing” effect can be created by combining le premier retour with a preposition (lors du premier retour ‘on the first return’); Section 4 will show that frame constituents are particularly prone to combine with introductory discourse particles, so-called ‘thematic markers’ (such as French quant à, à propos de etc. ‘as for’).

The second functional type of frame constituents is a subtype of HTs not often discussed in the syntactic literature (but see e.g., Fradin, 1990, “détachement sans rappel”; Jacobs, 2001, pp. 670–671 for German; Lambrecht, 2001, “unlinked topic”; Gergely Horváth, 2018, “topique libre” as opposed to “topique suspendue”, a loan translation of the syntactic notion of the HT described in Section 1). While syntactically it behaves like a frame, it denotes topical referents, that is, discourse referents about which the following predication holds:


Nos loisirs (‘our hobbies’) clearly constitutes the referent about which the following predication holds (‘our hobbies consist in going to concerts’), whereas le premier retour in (17) indicates the moment, that is, the temporal frame in which the proposition is integrated, the moment at which it is/was valid.

These two types of frames (frames with an adverbial function in a broad sense vs. isolated HTs in frame constructions) can be found in many Romance languages, while a third subtype of frame is only attested in French, to the best of our knowledge, where it is found relatively frequently in spontaneous dialogical data (cf. Gergely Horváth, 2018, p. 223: more than half of the frame constituents identified in a subpart of the Phonologie du français contemporain [PFC] corpus8 [see Durand et al., 2002] are of the TDc’est type). Stark (1997) called this type “TDc’est” (= “thème détaché c’est”, cf. Stark, 1997, chapter 2.2.b) and defined it as follows:

Ces structures seront appelées TDc’est et manifestent un degré d’attachement syntaxique supérieur à celui des TD. Elles représentent toujours des structures coordonnées au contexte précédent, être remplaçant le verbe principal, cela se référant à une action ou un état de choses (souvent l’actuel ‘sujet de séquence’) mentionnés auparavant:9


Curiously, these topical frames, despite the fact that they cannot be reinserted in the following clause and cannot have resumptive elements inside it, are prosodically (often no prosodic break at all) and syntactically more tightly bound to it than syntactic HTs: They are island-sensitive and acceptable in front of subordinate clauses (cf. Stark, 1997, chapter 2.2.b). Section 3 will discuss the fact that these properties seem to correlate with a higher “givenness” of the respective discourse referents in TDc’est constructions than for regular HTs or adverbial-like frame constituents.

To sum up, frame constituents can be defined as follows in a formal perspective (see also Kuong, 2006, p. 260, for a crosslinguistic perspective):


As we have seen, this formal definition covers at least two different functional constructions: (1) topic-marking constructions (HTs and French TDc’est) and (2) ‘frame setters’ with a domain-indication function in Krifka’s (2007) sense.

2.3 Syntactic Analyses

HTs and frames as defined in Sections 2.1 and 2.2 clearly show affinities with what has come to be known as the “(split) left periphery” since Rizzi’s (1997) seminal paper:


The “left periphery” is considered to be the place in the hierarchical sequence of functional projections constituting sentences where the clausal proposition is linked to the discursive context (either preceding or situational) and where operations such as topic-marking, focus-marking and indication of the illocutionary force of the respective utterance take place (explicitly by overt complementizers with finite clauses in Force° or with infinite clauses in Fin°, interrogative particles in Foc° - or in newer specifically interrogative functional heads). Rizzi identified two topical projections or “topical fields” (an important innovation by Benincà & Poletto, 2004 assuming a whole set of syntactically and semantically similar functional elements in TopP and FocP): a lower one associated with given, highly accessible topical referents (but see Benincà, 2001, pp. 56–57, with evidence from Italian ClLD against the lower TopP), and a higher one dedicated to topic-shift and similar operations (newer, inferable, but not already given topical referents). Frascarelli (2007) and Frascarelli and Hinterhölzl (2007) extended this observation, arguing on the empirical ground of Italian data, like Rizzi, and introduced three topic types: the “Aboutness-Shift Topic” (AS-Topic), the “Contrastive Topic” (C-Topic) and the “Familiar Topic” (Fam-Topic) (see article on “Topicalization in the Romance Languages”, in this encyclopedia). Examples 23 and 24 illustrate the first two topic types in Italian and French, while 25 illustrates a Fam-Topic:




(23) establishes a new topic or reintroduces (‘reactivates’) a topic in the discourse; (24) introduces an alternative topic (to the current one); and (25) mentions a given, usually unstressed topical discourse referent (often pronominal in nature). These three topic types are located in the left periphery as follows and are syntactically considered to be ClLD constituents (cf. Frascarelli, 2007, p. 698):


2.3.1 Hanging Topics

Hanging topics (HTs) are thus not included in the three functional topical projections that Frascarelli (2007) assumes:

They [= HT, ES] precede all other Topics (providing a sort of a “title”), with no syntactic connection with the rest of the sentence. . . . Hanging Topics have distinct formal and discourse properties with respect to the Aboutness-Shift Topic and are located in a specific (higher) position in the C-domain.

This position might be the specifier of what Benincà (2001) termed “DiscP”. Benincà (2001, p. 54) showed that, at least in Italian with some verbs, HTs may precede the complementizer che (‘that’) located in Force°:


This word order is not attested in French to the best of our knowledge, and it is impossible in all Romance languages for ClLD constituents (showing all kinds of connectivity effects such as those presented in Section 2.1).

Because of these and other observations on Paduan and Italian facts, Benincà (2001, p. 62) proposed the following revision of Rizzi’s (1997) model of the left periphery:


Crucially, “DiscP” is located above ForceP, probably hosting HTs such as questo libro (‘this book’) in 27. Bianchi and Frascarelli (2010, p. 77) presented similar data from Italian and concluded that, as the HT constituent may appear in initial position independent of the illocutionary force or type of the following utterance, these constituents do not belong syntactically to the following utterance, but are (coordinated?) discourse units on their own.

Data from Occitan complicate the picture: Here, HTs in embedded contexts preceding the complementizer and that are coreferential with an argument inside the embedded clause can be found (“troisième topicalisation”, Faure & Olivieri, 2013, p. 231):




These structures have been described and discussed by Camproux (1958); Lafont (1967); Sauzet (1989), and Lahne (2005). Faure and Olivieri (2013) summarized preceding analyses (especially Lahne, 2005) and pointed to the following properties of this construction: They are available in embedded contexts, like ClLD constituents (but see De Cat’s, 2007a, 2007b observations on French). However, as they do not agree in case with their resumptive elements, are always nominal in nature, are insensitive to syntactic islands, and do not show any reconstruction effects (see the introductory words to (12)), they are considered to be HTs (cf. (32) and (33) for the possible coreference of aquel òme with pro, which would be impossible if l’ostal d’aquel óme were moved from inside the embedded clause as in ClLD constructions; in (32), Old Occitan shows nominative marking on the HT element which is coreferential with an accusative gap in the following sentence):



The surprising word order HT que proposition, available also in embedded contexts, can be accounted for by considering introductory verbs such as sabi (‘I know’) in 32 as some sort of extrasyntactic parenthetical or “orphan” (see below; cf. Lahne, 2005).12 Alternatively, Faure and Olivieri (2013, pp. 261–265) proposed to analyze the HT in SpecForceP and que in (a strong)13 Force°, following Benincà (2001), but doing away with the assumed DiscP. As for its discourse function, the determiner phrase (DP) preceding que in Occitan is clearly the topic of the following sentence (Faure & Olivieri, 2013, p. 258, and chapter 3).

2.3.2 Frames

Differently from the Italian facts discussed so far for HT, French (temporal) frames such as la fameuse tempête (‘the famous storm’) in (34) and frame-like HTs such as le suisse-allemand (‘Swiss German’) in (35) always seem to follow overt complementizers (cf. Stark, 1997, p. 95):



(34) is structured as follows:

[ForceP il y a des jours [ForceP [TopP la fameuse tempête [TP pendant cinq heures [TP j’ai dû déblayer la neige]]]]]

That is, the assumption of DiscP14 above ForceP is not necessary or plausible for the French data.

In a discussion of medieval Romance data, Benincà (2006) refined her model and proposed the following sequence of functional heads in the left periphery:


Most relevant for the discussion is Benincà’s (2006, p. 56)15 observation that in modern Italian, adverbials, ‘scene setters’ (= frames) and (frame-like) HTs occupy the FrameP field (with frames/scene setters preceding HTs—but see Benincà and Poletto (2004) with the alternative assumption of the reversed order in Italian and Rhaeto-Romance) and the complementizers potentially in Frame° rather than Force°, in order to account for examples such as (25), contrary to (different types of) ClLD constituents, which are found in the TopP field.

Data from medieval Romance, especially the enclitic position of clitics after a verb in main clauses where the verb is not strictly in absolute initial position, corroborate Benincà’s (2006) view on Romance V2 as verb movement to a position not higher than FocP, with possible preceding HTs (37) and frame-like adverbials (38):



de Andrade (2018) discusses a specific type of HT in European Portuguese and Galician, where a full determiner phrase (DP) is preposed to a sentence containing both a demonstrative pronoun and a coreferential (also zero in Portuguese) clitic element:


He shows convincingly that this construction is available (although rarely attested in written sources) from the 16th century onward (de Andrade, 2018, p. 86; see below, Section 4) and that the DP element is located in FrameP, whereas the resumptive demonstrative pronoun sits in a left-peripheral, though lower topic position. (European) Portuguese and Galician thus seem to possess a particular frame-like HT.

Concluding this section, we wish to point out two things. First, both HT and frames belong without any doubt to what has since Rizzi (1997) been called the “split left periphery”, with many analyses couched in the Generative framework positioning frames higher than HTs in the structure. Italian facts, where HT, but not frames, may precede overt complementizers, indicate to a reverse order (DisP for HT, ForceP and FrameP for frames), but that seems to be highly language specific and needs further discussion. Second, alternatively to the assumption of their position in some higher left-peripheral position, such strongly disconnected constituents as HTs and frames can be seen as orphans, that is, phrases that are not integrated into sentences but form proper discourse constituents on their own (see e.g., Vallduví, 2002, for Catalan; Shaer & Frey, 2004, for German and English; Giorgi, 2015 for Italian), or more generally as constituents not integrated in, but preceding the following utterance (cf. Ott, 2015).

3. Functional Description: Topic-Marking (Shift?) and Framing

3.1 Hanging Topics (Function)

Ever since Barnes (1985), most of the specialized studies on Romance hanging topics (HTs) have been unanimous in attributing a topic-shift function to them (see e.g., Cinque, 1997, p. 95, for Italian; Pekarek-Doehler & Stoenica, 2012, p. 121, for French).16 The discussion in the literature suffers, however, from the ambiguous use of the term hanging topic (HT) discussed in Sections 1 and 2: Some authors use it syntactically (see (14) for our definition), others functionally. Maybe for that reason and using the term HT functionally, Frascarelli (cf. Frascarelli, 2007, p. 698, fn. 13) disagrees with this unidimensional functional analysis of HTs: In contrast to her “Aboutness-shift topic” (always realized as clitic left-dislocation, ClLD), HTs are, according to her, located higher in the structure and not necessarily associated with topic shift. This is confirmed by the empirical study of Stark (1997; see also Pekarek-Doehler et al., 2015), based on spontaneous conversational data from French: syntactically “frame-like” HTs such as those illustrated below (as well as “TDc’est” constituents), that is, syntactic frame constituents, may take up a given discourse topic, especially at the end of conversational units, in order to conclude:


Oysters represent the discourse topic for about 30 turns at least, and H sums up the advice he has given to X (how to recognize fresh vs. rotten oysters) before she recounted suffering from a bad dish, in order to reestablish the discourse topic from his conversational point of view (cf. Stark, 1997, p. 186).

3.2 Frames

The following example from Pekarek-Doehler and Stoenica (2012, p. 121) nicely illustrates the “shift function” of syntactic frames:


The authors state that after a short break, the nominal l’italien (‘Italian’) constitutes a kind of ‘interpretative frame’ mentioned by the speaker LIN, belonging ontologically to the same type of discourse referents as the prementioned l’anglais (‘English’). It is not commented on at this moment in the interaction, but rather later as the topic of line 09 (l’italien je le sais, ‘Italian, I know it’). Interestingly, the first mention of l’italien (‘Italian’) is realized also as a syntactic frame constituent, while the following sentence, which is about Italian, presents l’italien as a ClLD constituent.

Frame constituents in the sense of Krifka’s (2007) “frame setters” bear their domain-indicating function in their name: They denote or delimitate the interpretational frame for the following utterance without constituting its topic in a narrow sense (see above, Section 1, and e.g., Pekarek-Doehler et al., 2015, p. 64 for French; following Barnes, 1985; Lambrecht, 2001; Le Querler, 2003; Stark, 1997, to name but a few). Sometimes, metonymic relations between the frame constituent and an element in the following utterance can be identified (Pekarek-Doehler et al., 2015, p. 64; Stark, 1997, p. 91; cf. example 42), but there is no systematicity to be observed in the data (cf. example 43):



Pronominal frames, that is, strong pronouns as frame constituents, can have two different functions: One indicating an ‘individual frame’ (cf. Downing, 1991), and the other one being directly linked to turn aking:


Del has difficulties in being heard by the conversational group and indicates her desire to contribute to the discussion by repeating two pronominal frames, without however establishing herself as the notional frame of what follows, only as the one who is willing to speak.

The delimitation function of frames (see also Büring, 2016, p. 85), originally pointed out by Chafe (1976), was taken up by Charolles (1997, 2003) in an explicit textual or discursive coherence-oriented perspective for sentence-initial French adverbials and adverbial-like constituents, which are called “cadres de discours” and realized by “expressions cadratives” in Charolles’ work:

D’une manière différente, Charolles (1997, p. 4) propose de définir les cadres de discours comme des unités textuelles regroupant des propositions entretenant un même rapport avec un critère sémantique spécifié par un introducteur de cadre: celui-ci peut être d’ordre thématique, organisationnel, spatial, temporel, énonciatif . . . . (Prévost, 2003, p. 53)[In a different approach, Charolles (1997, p. 4) proposes to define discourse frames as textual units that share the same relation with a semantic feature signalled by a discourse marker: this can be a thematic, organizational, spatial, temporal or pragmatic feature.

(Prévost 2003, p. 53)]

Charolles is interested in his work in the textual “scope” of these discourse entities, which he considers as text-articulating reference points, as all the following utterances, until the establishment of a new frame, are to be interpreted with respect to the current frame. As ‘with respect to’ is a semantic relation close to ‘as for’, Charolles (2003), just like Prévost (2003) in a historical perspective on French, assumes a continuum between HTs and frames (cf. Prévost, 2003, p. 55, for details), which might be acceptable for our ‘hanging topic/frame’ type in Table 3.17 In contrast to temporal or spatial frame setters, ‘notional’, ‘individual frames’ à la Chafe are in fact very close to (hanging) topics:


(45a) and (45b) show adverbial constituents in initial position, while (45c) might establish Paul as the individual frame for what follows just as plausibly as signaling by this constituent the topic the following utterance is about. (45a) and (45b) could be reformulated by integrating the initial constituents (perhaps located not very high in the left periphery), which is excluded for Paul in (45c), which is thus syntactically a frame constituent. These examples illustrate the exclusively functional-textual approach of Charolles and his followers, confounding at times different constructions with different syntactic behaviors and constraints. Charolles et al. (2005, p. 125) note, however, that only autonomous, nonintegrated constituents such as our syntactic frames are prone to function as discourse frames (cf. Carruthers & LeDraoulec, 2017, p. 330):

In addition to “classical” approaches, which all deal with connection between (groups of) clauses, we present another approach: Discourse Framing. A discourse frame is described as the grouping together of a number of sentences which are linked by the fact that they must be interpreted with reference to a specific criterion, realised in an initial introducing expression. We suggest a distinction between two major modes of discourse organisation: whereas connection looks backwards toward previous text, discourse framing, or forward-labelling, looks ahead and provides instructions for the interpretation of forthcoming text. Within the discourse framing perspective, the asymmetry noted earlier between time and space seems to subside to give way to similar structuring roles for both dimensions.

Different sentence-initial adverbials have a different “frame potential”; in Charolles and Vigier (2005, p. 28), spatial and temporal scene setters are mentioned as prototypical frames, besides “adverbiaux praxéologiques” (e.g., en chimie, ‘in chemistry’), “médiatifs” (e.g., selon X, ‘according to X’) or “représentatifs” (e.g., dans le film de X, ‘in the movie by X’). Evaluative or speech-act-referring adverbials (which occupy a very high position in the left periphery) such as honnêtement (‘honestly’) or malheureusement (‘unfortunately’) are, in contrast, almost never seen in a discourse-framing function.

Besides the above-mentioned very general correlation between form and function (syntactic nonintegration = the possibility of constituting a discourse frame), Fuchs and Fournier (2003) observed that S-V order is necessary for the initial adverbial to constitute a discourse frame (something which also holds for HTs, see Lahousse, 2003 on French), whereas Carruthers and Le Draoulec (2017) found counterexamples in French fairy tales, albeit only with unaccusative verbs (where the internal argument is possible in postverbal position on independent grounds):


Register-oriented research on this kind of ‘discourse frame’ has been done by Carruthers (2011) on oral narratives and by Charolles and Pietrandrea (2012) on specific ‘frame markers’ such as en réalité in written French data.

3.3 Discourse-Pragmatic Status of HTs and Frames

The discourse-pragmatic status of HT referents is often described as “new” (cf. e.g., Cinque, 1997, p. 95). Following Ellen Prince’s work, (at least) “brand-new” and “unused” discourse referents are to be distinguished, which can be opposed to evoked and inferrable (“given”) ones (cf. Prince, 1981, 1992). While HT referents are often of the former kind, Stark (1997, chapter 5.1) also describes “given” discourse referents in frame-like HTs (see above, example 39), and identifies the topic-marking of given discourse referents and current discourse topics as the protoypical function of her “TDc’est” constructions (Stark, 1997, chapter 5.1.c). Definiteness and/or specificity are licit referential properties of HT or frame nominals; nonspecific referents and single quantifiers are excluded in both functions, as they are unable to establish topics or to indicate frames (see e.g., De Cat, 2007b, p. 504; Kiss, 2016, pp. 667–668).

Just like HTs, frame referents may thus either be “new” or “given” (cf. Stark, 1997, chapter V.2, of course, pronominal HTs or frames are always at least situationally given).

Stark (1997, chapter 5; see also Gergely Horváth, 2018, pp. 232–241) observed, in her conversational French data, a general correlation between the discourse-pragmatic status of HT and frame referents, their “persistence” in Givón’s (1992) terms, that is the scope of their topic or frame over subsequent discourse units, and their structural marking, especially their prosodic marking (see also Hamlaoui & Szendröi, 2015, for similar observations in Hungarian and Básáà):


Given HT or frame referents promoted to topic (= topic shift) or “frame setter” with scope only over the respective utterance do not have thematic markers (see below, Section 4) and have no prominent intonational profile.


Given HT or frame referents promoted to topic (= topic shift) or “frame setter” with a higher “persistence”, instead, may sometimes have thematic markers and are marked by a higher rising tone and often a following pause.


Given HT or frame referents which already constitute the current topic or frame are marked like the first group and often sum up the speaker’s opinion on a certain subject (see example 40 above). In cases where the speaker reformulates her own preceding statements, avec is usually found in French with frames of this type:


New HT or frame referents promoted to topic (= topic shift) or “frame setter” with scope only over the respective utterance are marked by their own intonational contour (IC); frames are often also marked by a falling tone and a following pause plus some specific thematic markers (such as point de vue, chez or pour):

Moi clearly represents a Contrastive Topic in this utterance but is not taken up in the subsequent conversation.


New HT or frame referents promoted to topic (= topic shift) or “frame setter” with a considerable scope are always marked by their own IC, often a falling intonation plus a pause, and very systematically by thematic markers such as en ce qui concerne, sur le plan de (‘as for’) and the like:

To sum up, HT and frames mark topics and domains relevant for the following utterance(s), with a certain preference for new discourse referents in these functions (“shift”). There is a correlation, at least in naturalistic French data, between the discourse-pragmatic status of the respective referents and their prosodic and categorical marking: The newer the discourse referent, the more prominent its intonational profile and the more likely the presence of thematic markers.

The following section will take a closer look at these elements as well as at their coming into being and the general diachrony of HT and frame constructions in Romance.

4. A Brief Glance at Diachrony, Markedness, and Thematic Markers

Under the assumption of one or two universally available and syntactically independent initial positions in utterances dedicated to delimiting the domain in which the following predication(s) holds and/or to establishing a (new) sentence or discourse topic, hanging topics (HTs) and frames would not be expected to exhibit true diachrony in the sense of changing forms and functions. This position is actually held in Stark (1999b, pp. 142–144) and Halla-aho (2016) (especially on pp. 388–389; see also Bortolussi, 2017 on the colloquial status of topicalization constructions in Latin), both of whom only acknowledge changes in the variational markedness of the single constructions as well as of different thematic markers as potential diachronic phenomena, most frequently the simple acceptability in standard varieties vs. a vernacular or oral flavor:

Lediglich in dem eng begrenzten Bereich der Einleitungsfloskeln lässt sich Sprachwandel als beständige Erneuerung des Paradigmas konstatieren, nicht aber beim Auftreten von Voranstellungsstrukturen im allgemeinen, die als möglicherweise universale Versprachlichungsstrategien keine Geschichte haben.

(Stark, 1999b, p. 144)18

In (18), repeated here for convenience, it was shown that it is only the lexical choice (retour, ‘return’, vs. fois, ‘time’) and not the syntactic construction in itself which distinguishes a marked frame constituent from a garden-variety initial adverbial:


4.1 Hanging Topics (and Clitic Left-Dislocation) and Frames

In fact, HT and frames have always been available in Romance as well as in Latin, where they are known as instances of nominativus and less frequently also as (adverbial) accusativus pendens (cf. e.g., Faure & Olivieri, 2013, p. 252; see also Müller-Lancé, 1994, on “absolute constructions”, functionally very close to frames, or Halla-aho, 2016, pp. 367, 388–389 on Latin left-dislocations for example in Cato’s, Plautus’s or Terence’s works, with some examples from Cicero and many from Christian Latin, sometimes with a case difference between the preposed topic and the resumptive pronoun, thus satisfying one criterion for an HT classification; see also Bortolussi, 2017 for more HT and clitic left-disclocation [ClLD] examples):


Though very infrequent in older texts,19 frame constituents like these are attested throughout the centuries, but are usually considered to be nonstandard or “oral” in nature (Stark, 1997, chapter 6, proposed considering them as medium-dependent discourse structuring devices that are not necessary in written communication).

Bouzouita (2014), as mentioned in Section 2, provides examples of HT constructions with a resumptive element in 13th15th-century Spanish, such as the following one:


That left-peripheral topicalization structures are most probably panchronic and universal in nature is evident from medieval Romance sources such as the Placiti cassinesi, around 960, one of the first attestations of Old Italian:


The nominal subject constituent [kelle terre . . .]DP is taken up by le in the following clause, of which it clearly constitutes the (contrastive?) topic—a case of ClLD.

The very few observations on Romanian HTs (also called “hanging themes” in the Romanian literature) come from Rodica Zafiu’s (2013, see below) as well as Gabriela Pană Dindelegan’s (2016) work, including some examples from 16th-century Romanian, such as the following. This example includes a proper name in the left periphery, taken up by a subject pro in the following conditional clause and a dative pronoun plus dative clitic in the root clause:


de Andrade (2018) also shows structural identity between HT and frame-like HT constructions in Portuguese (and Galician) at least from the 16th century onward, and explicitly excludes HT constructions from linguistic change involving ClLD and topicalization structures in the history of Portuguese.

4.2 The Role of Thematic Markers with HTs and Frames

Initial HT or frame constituents seem to be more standard-like when introduced by thematic markers, which transforms them into prepositional structures in a broad sense (cf. Stark, 1999b, p. 135). This can be illustrated with example (55):


Combining à with the frame constituent yields a regular spatial frame adverbial à l’intérieur (‘from inside’).

This observation also holds for the whole paradigm of thematic markers such as à propos de, en cuanto a, per quanto riguarda, Lat. de, quantum as (‘as for’; see Bortolussi, 2017, pp. 105–107, on Latin topicalization structures with explicit marking); that is, prepositional or univerbalized preposition-like monovalent elements (sometimes even containing a finite verb, cf. Fr. en ce qui concerne X, literally ‘in that which concerns X’), which explicitly signal the topic-marking or frame-marking function of HTs and frames and correlate in frequency with the ‘new’ and persistent character of the respective HT and frame discourse referents (see e.g., Morel, 1992, for French). These elements can also be found in Romanian (cât despre, ‘as for’, dacă e vorba de, ‘as we speak about’) and are described there as typical HT marking devices (cf. Zafiu, 2013, p. 573):


Following work by Noailly-Le Bihan (1982), Stark (1999b, p. 137) described potential changes in the lexical paradigm of French thematic markers, where younger elements such as du côté de, au niveau (de), and so on are reduced to their mere lexical core in oral varieties, resulting in the following examples:


Nouns like côté (‘side’) or question (‘question’) lose their nominal character in such contexts and turn into prepositions; semantically, most originally ‘local’ thematic markers have lost their original semantics (e.g., au niveau de or sur le plan de, ‘at the level of’) and can be considered to mean ‘as for’. Thematic markers containing a speech-act or matter-of-fact-referring element such as Fr. propos (‘proposal’, ‘topic’), matière (‘matter’, ‘topic’), question (‘question’), and the like do not undergo any substantial semantic change and represent something like the core of thematic markers.

Combettes (2003) presented valuable insights into the diachrony of thematic markers in French, postulating a general path of change via reanalysis from postverbal argumental or adjunct elements inside the verb phrase (VP) toward peripheral topic or frame markers. For example, au regard de in (58) is a regular preposition introducing the complement of amour (‘love’):


Two centuries later, au regard de is attested as a thematic marker (with “portée énonciative” in Combettes’ terms) with a postverbal adjunct (also attested by then in initial position) assuming thematic framing:


The key ingredient in such reanalyses is, according to Combettes (2003, pp. 144, 146), the introduction of topical elements by the respective prepositions, which then may be reanalyzed syntactically as occupying peripheral positions (HT or frame). Besides, Combettes (2003, p. 151) observed the presence of frame constituents in medieval and preclassical Latin and French texts with quantum ad (quant à) with a “paragraph introducing function”, which corresponds exactly to type v in Section 3.

Functionally, after having given a very good overview of the abundant research into single thematic markers in French (see, e.g., Anscombre, 2006; Noailly, 2006; Porhiel, 2004), Velghe and Lahousse (2015) investigated three thematic markers in French and their cooccurrence with HTs and frames, and found, based on corpus data, that pour ce qui est de is more prone to mark (new) aboutness topics and HTs, whereas (au) niveau (de) and en matière de (this marker is analyzed for the first time) are more often found with frame constituents (which correspond also to the semantics of their core noun).20

5. Conclusion

This overview of hanging topics (HT) and frames in Romance languages, that is, on two left-peripheral constructions that may have been universally available and attested since preclassical Latin (e.g., in Plautus), revealed two main points.

First, the notion of HT is polysemous and used in at least two different readings in the specialized literature: formally, as a specific syntactic construction (left-peripheral non-case-marked nominals without any connectivity effects with the following sentence and possibly taken up there by a resumptive element); or functionally, as a relatively independent topic-marking (often topic-shifting) construction. The notion of frame is polysemous as well, as it might, on the formal side, denote completely independent constituents in front of complete sentences where they cannot be taken up by any kind of resumptive element (these constituents may denote either the topic or the referential domain the following proposition is connected to); or, functionally, the domain in which the following proposition is valid. Both constructions are attested throughout the whole documented history of Romance languages and may differ in their variationist markedness or status (being most often considered, however, to indicate a colloquial, oral register), in many cases depending on the presence or absence of so-called thematic markers.

Second, research on HT and frames in different Romance languages focuses, curiously, on quite different aspects so that a thorough comparative study is still lacking to date. Work on French HTs and frames is often functionalist and variationist in nature (with the clear exception of De Cat) and tries to pin down the precise functions of HT and frames in discursive interactions, whereas studies on HTs and frames (and clitic left-dislocation, or ClLD) in Italoromance focus very much on formal syntactic analyses in what it has become customary to refer to as the “split left periphery”. Portuguese and Galician in particular are to some extent underresearched as to these two constructions, and work on Ibero-Romance (especially on Catalan) has in general a neat focus on prosodic properties of left-peripheral constituents and/or tends to exclude frames from the picture and center around ClLD vs. HT constructions. This last remark also holds for the few observations on Romanian we found (only HT), and minor Romance languages (except Occitan) are often not discussed with respect to HT and frames. The most important lacuna in the field is without any doubt the lack of a thorough comparative and if possible Pan-Romance study on the prosodic, syntactic, and functional characteristics of HT and frames.


  • Alexiadou, A. (2006). Left dislocation (including CLLD). In M. Everaert, H. Van Riemsdjik, R. Goedemans, & B. Hollebrandse (Eds.), The Blackwell companion to syntax (19th ed., pp. 668–700). Oxford, UK: Blackwell.
  • de Andrade, A. (2018). A special type of left dislocation in Galician and Portuguese: The D-construction. Estudos de Linguistica Galega. Volume especial I, 81–105.
  • Ariel, M. (1991). The function of accessibility in a theory of grammar. Journal of Pragmatics, 16(5), 443–463.
  • Barnes, B. K. (1985). The pragmatics of left detachment in spoken standard French. Amsterdam, The Netherlands: John Benjamins.
  • Baumann, S, & Riester, A. (2012). Referential and lexical givenness: Semantic, prosodic and cognitive aspects. In G. Elordieta & P. Prieto (Eds.), Prosody and meaning (pp. 119–161). Berlin, Germany: De Gruyter Mouton.
  • Bednarek, M. A. (2005). Frames revisited—The coherence-inducing function of frames. Journal of Pragmatics, 37, 685–705.
  • Benincà, P. (2001). The position of topic and focus in the left periphery. In C. Guglielmo & S. Giampaolo (Eds.), Current studies in Italian syntax: Essays offered to Lorenzo Renzi (59th ed., pp. 39–64). Amsterdam, The Netherlands: Elsevier-North Holland.
  • Benincà, P. (2006). A detailed map of the left periphery of medieval Romance. In R. Zanuttini, H. Campos, E. Herburger, & P. H. Portner (Eds.), Crosslinguistic research in syntax and semantics: Negation, tense, and clausal architecture (pp. 53–86). Washington, DC: Georgetown University Press.
  • Benincà, P., & Poletto, C. (2004). Topic, focus and V2: Defining the CP sublayers. In L. Rizzi (Ed.), The structure of CP and IP: The cartography of syntactic structures (2nd ed., pp. 52–75). Oxford, UK: Oxford University Press.
  • Bianchi, V., & Frascarelli, M. (2010). Is topic a root Phenomenon? Iberia, 2(1), 43–88.
  • Blackwell, S. E. (2017). Frames of references and antecedentless anaphora in Spanish conversation. Journal of Psycholinguist Research, 47(2), 283–395.
  • Bortolussi, B. (2017). Topicalizations, left dislocations and the left-periphery. Catalan Journal of Linguistics, 16, 101–123.
  • Bouzouita, M. (2014). Left dislocation phenomena in Old Spanish: An examination of their structural properties. In A. Dufter & A. S. O. de Toledo y Huerta (Eds.), Left sentence peripheries in Spanish: Diachronic, variationist and comparative Perspectives (pp. 23–51). Amsterdam, The Netherlands: John Benjamins.
  • Büring, D. (2016). (Contrastive) topic. In C. Féry & S. Ishihara (Eds.), The Oxford handbook of information structure (pp. 64–85). Oxford, UK: Oxford University Press.
  • Cadiot, P. (1978). “Où t’as ta femme ?” Semantikos, 2(3), pp. 1–20.
  • Camproux, C. (1958). Etude syntaxique des parlers gévaudanais. Paris, France: Presses Universitaires de France.
  • Carruthers, J. (2011). Temporal framing in the conte. From theoretical debate to oral story performance. Journal of French Language Studies, 65(4), 488–504.
  • Carruthers, J., & Le Draoulec, A. (2017). Encadrement discursif et postposition du sujet: l’éclairage des contes oraux. Journal of French Language Studies, 27, 329–353.
  • Casielles-Suárez, E. (2003). Left-dislocated structures in Spanish. Hispania, 86(2), 326–338.
  • Casielles-Suárez, E. (2004). The syntax-information structure interface: Evidence from Spanish and English. New York, NY: Routledge.
  • Chafe, W. L. (1976). Givenness, contrastiveness, definiteness, subjects, topics and point of view. In N. L. Charles (Ed.), Subject and topic (pp. 27–55). New York, NY: Academic Press.
  • Charolles, M. (1997). L’encadrement du discours: univers, champs, domaines et espaces. Monographie Cahier de Recherche Linguistique, LANDISCO, URA-CNRS 1035, Université de Nancy, 2(6), 1–73.
  • Charolles, M. (2003). De la topicalité des adverbiaux détachés en tête de phrase. In Travaux de Linguistique, 47(2), 11–51.
  • Charolles, M., Le Draoulec, A., Pery-Woodley, M. P., & Sarda, L. (2005). Temporal and spatial dimensions of discourse organisation. Journal of French Languages Studies, 15(2), 115–130.
  • Charolles, M., & Pietrandrea, P. (2012). En réalité: de la modalisation à l’organisation du discours. Travaux de linguistique, 64(1), 111–142.
  • Charolles, M., & Vigier, D. (2005). Les adverbiaux en position préverbale: portée cadrative et organisation des discours. Langue française, 148(4), 9–30.
  • Chomsky, N. (1977). On Wh-movement. In P. W. Culicover, T. Wasw, & A. Akmajian (Eds.) Formal syntax (pp. 71–133). San Francisco, CA: Academic Press.
  • Cinque, G. (1977). The movement nature of left-dislocation. Linguistic Inquiry, 8(2), 397–412.
  • Cinque, G. (1979). Left-dislocation in Italian: A syntactic and pragmatic analysis. Cahiers de lexicologie, 34, 96–127.
  • Cinque, G. (1997). ‘Topic’ constructions in some European languages and ‘connectedness’. In E. Agnostoupoulou, H. van Riemsdijk, & F. Zwarts (Eds.), Materials on left dislocation (pp. 93–118). Amsterdam, The Netherlands: Benjamins.
  • Combettes, B. (2003). Le topique comme constituent périphérique: aspects diachroniques. Travaux de linguistique, 47(2), 137–161.
  • De Cat, C. (2007a). French dislocations. Oxford, UK: Oxford University Press.
  • De Cat, C. (2007b). French dislocation without movement. Natural Language and Linguistic Theory, 25, 485–534.
  • Delais-Roussarie, E., & Feldhausen, I. (2014). Variation in prosodic boundary strength in French. In N. Campbell, D. Gibbon, & D. Hirst (Eds.), Proceedings of speech prosody 7 (pp. 1052–1056). Dublin.
  • Delais-Roussarie, E., Doetjes, J., & Sleeman, P. (2004). Dislocation. In F. Corblin & H. de Swart (Eds.). Handbook of French semantics (pp. 505–530). Palo Alto, CA: CSLI Publications.
  • Dolci, R. (1986). Algunas construcciones con anteposición de constituyentes oracionales en español: su determinación y análisis sintáctico (PhD dissertation, Università degli Studi di Venezia).
  • Downing, A. (1991). An alternative approach to theme: A systemic-functional perspective. Word, 42(2), 119–143.
  • Durand, J., Laks, B., & Lyche, C. (2002). La phonologie du français contemporain: usages, variétés et structure. In C. D. Pusch & W. Raible (Eds.), Romanistische Korpuslinguistik- Korpora und gesprochene Sprache/Romance corpus linguistics—Corpora and spoken language (126th ed., Scripta Oralia, pp. 93–106). Tübingen, Germany: Narr. (Corpus PFC)
  • Erteshik-Shir, N. (1997). The dynamics of focus structure. Cambridge, UK: Cambridge University Press.
  • Erteshik-Shir, N. (1999). Focus structure and scope. In G. Rebuschi & L. Tuller (Eds.), Grammar of focus (pp. 119–150). Amsterdam, The Netherlands: Benjamins.
  • Escobar, M. A. (1995). Lefthand satellites in Spanish. Utrecht, The Netherlands: OTS.
  • Faure, R., & Oliviéri, M. (2013). Stratégies de topicalisation en occitan. Corpus, 12, 231–270.
  • Feldhausen, I. (2010). Sentential form and prosodic structure in Catalan. Amsterdam, The Netherlands: Benjamins.
  • Feldhausen, I. (2014a). Intonation and preverbal subjects in Italian. Proceedings of the 10th International Seminar on Speech Production (ISSP) (pp. 118–121).
  • Feldhausen, I. (2014b). Modeling individual variation in prosody: The case of Spanish clitic left-dislocations. Proceedings of the 10th International Seminar on Speech Production (ISSP) (pp. 114–117).
  • Feldhausen, I. (2016a). The relation between prosody and syntax: The case of different types of left-dislocations in Spanish. In Meghan Armstrong, Nicholas Henriksen, & Maria del Mar Vanrell (Eds.), Intonational grammar in Ibero-Romance: Approaches across linguistic subfields (pp. 153–180). Amsterdam, The Netherlands: Benjamins.
  • Feldhausen, I. (2016b). Inter-speaker variation, optimality theory and the prosody of clitic left-dislocations in Spanish. Probus, 28(2), 293–334.
  • Feldhausen, I., & Lausecker, A. (2018). Diatopic variation in prosody: Left- and right-dislocations in Spanish. In M. Belz, S. Fuchs, S. Jannedy, C. Mooshammer, O. Rasskazova, & M. Zygis (Eds.), Proceedings of phonetics and phonology in the German-speaking countries (P&P 13), Berlin (Germany), September 2017 (pp. 57–60). Berlin, Germany: Humboldt University of Berlin.
  • Fontana, J. (1993). Phrase structure and the syntax of clitics in the history of Spanish (PhD dissertation, University of Pennsylvania).
  • Fradin, B. (1990). Approche des constructions à détachement: Inventaire. Revue Romane, 25(1), 3–34.
  • Frascarelli, M. (2000). The syntax–phonology interface in focus and topic constructions in Italian. Dordrecht, The Netherlands: Kluwer.
  • Frascarelli, M. (2007). Subjects, topics and the interpretation of referential ‘pro’. An interface approach to the linking of (null) pronouns. Natural Language and Linguistic Theory, 25, 691–734.
  • Frascarelli, M. (2017). Dislocations and framings. In A. Dufter & E. Stark (Eds.), Manual of Romance morphosyntax and syntax (pp. 472–501). Berlin, Germany: de Gruyter.
  • Frascarelli, M., & Hinterhölzl, R. (2007). Types of topics in German and Italian. In S. Winkler & K. Schwabe (Eds.), On information structure, meaning and form (pp. 87–116). Amsterdam, The Netherlands: John Benjamins.
  • Frota, S. (2000). Prosody and focus in European Portuguese: Phonological phrasing and intonation. New York, NY: Garland.
  • Fuchs, C., & Fournier, N. (2003). Du rôle cadratif des compléments localisants initiaux selon la position du sujet initial. Travaux de linguistique, 47, 79–109.
  • Gergely Horváth, M. (2018). Le français parlé informel: Stratégies de topicalisation. Berlin, Germany: de Gruyter.
  • Gili Fivela, B. (1999). The prosody of left-dislocated topic constituents in Italian read speech. In Proceedings of the sixth European conference on speech communication and technology—Eurospeech, Budapest (pp. 531–534).
  • Giorgi, A. (2015). Discourse and the syntax of the left periphery: Clitic left dislocation and hanging topic. In J. Bayer, R. Hinterhölzl, & A. Trotzke (Eds.). Discourse-oriented syntax (pp. 229–250). Amsterdam, The Netherlands: Benjamins.
  • Givón, T. (1992). The grammar of referential coherence as mental processing instructions. Linguistics, 30(1), 5–55.
  • Grohmann, K. (2000). Movement issues in left dislocation constructions. University of Pennsylvania Working Papers in Linguistics, 7(1), 75–86.
  • Gundel, J. K. (1975). Left dislocation and the role of topic-comment structure in linguistic theory. In Working papers in linguistics (Vol. 18, pp. 72–131). Columbus, OH: University Columbus Ohio.
  • Gundel, J. K. (1977). The role of topic and comment in linguistic theory. Austin: Indiana University Linguistics Club.
  • Haegeman, L. (2009). Parenthetical adverbials: The radical orphanage approach. In B. Shaer, P. Cook, W. Frey, & C. Maienborn (Eds.), Dislocated elements in discourse (pp. 331–347). New York, NY: Routledge. (Original work published 1991)
  • Haegeman, L., & Greco, C. (2018). Main clause external constituents and the derivation of subject initial verb second. Nederlandse taalkunde, 23(1), 23–55.
  • Halla-aho, H. (2016). Left-detached constructions from early to late Latin (nominatiuus pendens and attractio inuersa). In J. N. Adams & N. Vincent (Eds.). Early and late Latin: Continuity or change? (pp. 367–389). Cambridge, UK: Cambridge University Press.
  • Hamlaoui, F., & Szendröi, K. (2015). A flexible approach to the mapping of intonational phrases. Phonology, 32, 79–110.
  • Heim, I. R. (1982). The semantics of definite and indefinite noun phrase (PhD dissertation, University of Massachusetts).
  • Hoffmann, M. E. (1989). A typology of Latin theme constituents. Cahiers de l’Institut de linguistique de Louvain, 15(1–4), 185–196.
  • Jacobs, J. (2001). The dimensions of topic-comment. Linguistics, 39(4), 641–681.
  • Kamp, H., & Reyle, U. (1993). From discourse to logic. Dordrecht, The Netherlands: Kluwer.
  • Kiss, K. É. (2016). Discourse functions: The case of Hungarian. In C. Féry & S. Ishihara (Eds.). The Oxford handbook of information structure (pp. 663–685). Oxford, UK: Oxford University Press.
  • Krifka, M. (2007). Basic notions of information structure. In C. Fery & M. Krifka (Eds.), Interdisciplinary studies on information structure (6th ed., pp. 13–56). Potsdam, Germany: Universitätsverlag.
  • Kuong, I.-K. (2006). Clausal peripheries and resumptives: A cross-linguistic study of topic-comment structures (PhD dissertation, Georgetown University).
  • Lafont, R. (1967). La phrase occitane. Paris, France: Presses Universitaires de France.
  • Lahne, A. (2005). Die linke Satzperipherie im Okzitanischen [MA thesis, Friedrich Schiller- University, Jena].
  • Lahousse, K. (2003). La complexité de la notion de topique et l'inversion du sujet nominal. Travaux de linguistique, 47, 111–136.
  • Lambrecht, K. (2001). Dislocation. In M. Haspelmath, E. König, W. Oesterreicher, & W. Raible (Eds.). Language typology and language universa ls: An international handbook (Vol. 2, pp. 1050–1078). Berlin: de Gruyter.
  • Larsson, E. (1979). La dislocation en français. Etude de syntaxe générative. Lund, Sweden: LiberLäromedel Gleerup.
  • Le Querler, N. (2003). Le nominativus pendens en français. Cahiers de praxématique, 40, 149–166.
  • López, L. (2009). A derivational syntax for information structure. Oxford, UK: Oxford University Press.
  • López, L. (2016). Dislocations and information structure. In C. Féry & S. Ishihara (Eds.), The Oxford handbook of information structure (pp. 402–421). Oxford, UK: Oxford University Press.
  • Minsky, M. L. (1977). Frame-system theory. In P. N. Johnson-Laird & P. C. Watson (Eds.), Thinking: Readings in cognitive science (pp. 355–376). Cambridge, UK: Cambridge University Press.
  • Morel, M.-A. (1992). L’opposition thème/rhème dans la structuration des dialogues oraux. Journal of French Language Studies, 2(1), 61–74.
  • Müller-Lancé, J. (1994). Absolute Konstruktionen vom Altlatein bis zum Neufranzösischen: Ein Epochenvergleich unter Berücksichtigung von Mündlichkeit und Schriftlichkeit. Tübingen, Germany: Narr.
  • Nespor, M., & Vogel, I. (1986). Prosodic phonology. Dordrecht, The Netherlands: Foris.
  • Noailly, M. (2006). Quoi de neuf côté préposition?Modèles linguistiques, 27(53), 75–90.
  • Noailly-Le Bihan, M. (1982). Côté, question et quelques autres. Linguisticae Investigationes, 6(2), 333–343.
  • Ott, D. (2015). Connectivity in left-dislocation and the composition of the left periphery. Linguistic Variation, 15(2), 225–290.
  • Pablos, L. (2006). Pre-verbal structure building in Romance languages and Basque (PhD dissertation, University of Maryland).
  • Pekarek-Doehler, S., De Stefani, E., & Horlacher, A.-S. (2015). Time and emergence in grammar: Dislocation, topicalization, and hanging topic in French talk-in-interaction. Amsterdam, The Netherlands: Benjamins.
  • Pană Dindelegan, G. (ed.). (2016). The syntax of Old Romanian. Oxford, UK: Oxford University Press.
  • Poletto, C., & Bocci, G. (2016). Syntactic and prosodic effects of information structure in Romance. In C. Féry & S. Ishihara (Eds.), The Oxford handbook of information structure (pp. 642–662). Oxford, UK: Oxford University Press.
  • Porhiel, S. (2004). Les introducteurs de cadre thématique. Cahiers Lixicologiques, 85, 9–45.
  • Prévost, S. (2003). Les compléments spatiaux: du topique au focus en passant par les cadres. Travaux de linguistique, 47, 51–77.
  • Prince, E. (1981). Towards a taxonomy of given-new information. In P. Cole (Ed.), Radical pragmatics (pp. 223–256). New York, NY: Academic Press.
  • Prince, E. (1992). The ZPG letter: Subjects, definiteness, and information-status. In S. Thompson & W. Mann (Eds.), Discourse description: Diverse analyses of a fundraising text (pp. 295–325). Amsterdam, The Netherlands: Benjamins.
  • Reinhart, T. (1981). Pragmatics and linguistics: An analysis of sentence topics. Philosophica, 27, 53–94.
  • Rivero, M. L. (1980). On left-dislocation and topicalization in Spanish. Linguistic Inquiry, 11, 363–393.
  • Rizzi, L. (1997). The fine structure of the left periphery. In L. Haegeman (Ed.), Elements of grammar (pp. 281–337). Dordrecht, The Netherlands: Kluwer.
  • Ross, J. R. (1967). Constraints on variables in syntax (PhD dissertation, Massachusetts Institute of Technology).
  • Rossi, M. (1999). L’Intonation, le Système du Français - Description et Modélisation. Paris, France: Ophrys.
  • Sauzet, P. (1989). Topicalisation et prolepse en occitan. Revue des langues romanes, 93(2), 235–273.
  • Schank, R. C., & Abelson, R. P. (1977). Scripts, plans, goals and understanding: An inquiry into human knowledge structures. New York, NY: L. Erlbaum.
  • Selkirk, E. (1984). Phonology and syntax: The relation between sound and structure. Cambridge, MA: MIT Press.
  • Shaer, B., & Frey, W. (2004). ‘Integrated’ and ‘non integrated’ left peripheral elements in German and English. In B. Shaer, W. Frey, & C. Maienborn (Eds.), Dislocated elements workshop (November 2003, ZAS Papers in Linguistics 35, pp. 465–502). Berlin, Germany: Zentrum für Allgemeine Sprachwissenschaft.
  • Silva Corvalán, C. (1984). Topicalización y pragmática en español. Revista Española de Lingüística, 14(1), 1–20.
  • Somers, M. H. (1994). Theme and topic: The relation between discourse and constituent fronting in Latin. In J. Herman (Ed.), Linguistic studies on Latin: Selected papers from the 6th international colloquium on Latin linguistics (Budapest, 23–27 March 1991) (pp. 151–163). Amsterdam, The Netherlands: Benjamins.
  • Stark, E. (1997). Voranstellungsstrukturen und topic-Markierung im Französischen. Mit einem Ausblick auf das Italienische. Tübingen, Germany: Narr.
  • Stark, E. (1999a). Antéposition et marquage du thème (topic) dans les dialogues spontanés. In C. Guimier (Ed.), La thématisation dans les langues. Actes du colloque de Caen 9–11 octobre 1997 (pp. 337–358). Bern, Switzerland: Lang.
  • Stark, E. (1999b). Französische Voranstellungsstrukturen—Grammatikalisierung oder universale Diskursstrategien? In J. Lang & I. Neumann-Holzschuh (Eds.), Reanalyse und Grammatikalisierung in den romanischen Sprachen (pp. 129–146). Tübingen, Germany: Niemeyer.
  • Stark, E., Ueberwasser, S., & Göhring, A. (2014–2020). KorpusWhat’s up, Switzerland?” Zurich, Switzerland: University of Zurich.
  • Tannen, D. (1979). What’s in a frame. Surface evidence for underlying expectations. In R. O. Freedle (Ed.), New directions in discourse processing (pp. 137–181). Nordwood, NJ: Ablex.
  • Vallduví, E. (1992). The informational component. New York, NY: Garland.
  • Vallduví, E. (2002). L’oració com a unitat informativa. In J. Solà, M.-R. Lloret, J. Mascarò, & M. P. Saldanya (Eds.), Gramàtica del català contemorani (pp. 1221–1279). Barcelona, Spain: Empúries.
  • Velghe, T., & Lahousse, K. (2015). Thematic markers in informal written French: pour ce qui est de, au niveau (de), and en matière de. Journal of French Language Studies, 25(3), 423–444.
  • Villalba, X. (2000). The syntax of sentence periphery (Unpublished doctoral dissertation, Universitat Autonoma de Barcelona).
  • Villalba, X. (2009). The syntax and semantics of dislocations in Catalan: A study of asymmetric syntax at the peripheries of sentence. Cologne, Germany: Lambert Academic Publishing.
  • Zafiu, Rodica (2013). Information structure. In G. P. Dindelegan & M. Maiden (Eds.), The Grammar of Romanian (pp. 568–574). Oxford: Oxford University Press.
  • Zubizarreta, M. L. (1998). Prosody, focus and word order. Cambridge, MA: MIT Press.


  • 1. See Stark (1997), chapter 1.4., for an overview of the history of the notion of topic and frame in information structure terms. Erteshik-Shir (1997, 1999) calls the latter elements “topiques scéniques.”

  • 2. Sometimes, hanging topics at the beginning of utterances are also called hanging topic left dislocations (HTLDs), e.g., in Cinque (1997) or De Cat (2007a, 2007b), but we will stick to HT in this chapter (see also López’s 2016 term, “H-type dislocations“).

  • 3. For more details on the distinction between HT and CllD in Spanish see, for example, Casielles-Suárez (2003, 2004); Escobar (1995); Feldhausen (2016a); Villalba (2000); Zubizarreta (1998).

  • 4.

  • 5. Note that, at least for French, there is no agreement about the compatibility of HT constructions with resumptive clitics (Delais-Roussarie et al., 2004 vs. Grohmann, 2000); for Italian, Benincà (2006, p. 57), and Frascarelli (2007, p. 66) even assume an obligatory resumptive element in the following clause. Usually, resumptive clitics are considered to be a characteristic of ClLD and ClRD (cf. Alexiadou, 2006). HTs always precede ClLD constituents.

  • 8. ‘These structures will be denoted TDc'est and they exhibit a higher degree of syntactic connectedness than TD. They are always coordinated with the preceding context, with être replacing the main verb, and cela referring to a previously mentioned action or state of affairs (often the current topic).’

  • 9. But see Gergely Horváth (2018, p. 233) for different results.

  • 12. This is assuming a parametrization for Force°, which in the strong case in declaratives always has to be spelled out and always provokes movement of a constituent to its specifier, which could account for the different ordering of HT constituents and complementizers in French and Occitan.

  • 13. A plausible derivation of many Germanic V3 orders, especially in West Flemish, most certainly needs the assumption of FrameP above ForceP (cf. Haegeman & Greco, 2018).

  • 14. See also Benincà and Poletto (2004) with evidence from contemporary standard Italian data.

  • 15. But see Benincà (2001, p. 44) who mainly claims a stylistic rather than a functional difference between ClLD and HT in Italian.

  • 16. However, neither Charolles nor Prévost are interested in syntactic details and describe the respective constructions almost exclusively on functional grounds.

  • 17. ‘Only in the narrow domain of thematic markers can a linguistic change in the sense of a continuous renewal of the paradigm be stated, but not as far as the occurrence of frames is concerned. As they are universal strategies of discourse organization, they cannot possibly have a proper history.’

  • 18. Cf. Hoffmann (1989), on Classical Latin; Somers (1994), on frame-like constituents in Cicero’s letters, mostly encoded as cases of ablativus absolutus; Prévost (2003), on 15th-century French; Stark (1997), chapter 6.3.b on Old French.

  • 19. Bibliographical research on hanging topics and frames in Romance linguistics did not yield results in the domain of “thematic markers” for other Romance languages.

  • 20. But see Pekarek-Doehler et al. (2015, p. 67) who also attribute frame setting to HT constituents.