https://orcid.org/0000-0001-7844-8946
Abstract
This paper empirically and theoretically investigates the syntax of split topicalization in Jordanian Arabic, a previously undiscovered phenomenon. The results of a large-scale acceptability judgment task (n = 463) reveal that Jordanian Arabic displays split topicalization of various categorial types. Building on the labeling framework, the remnant movement analysis and the idea that there exists an IP-internal focus position below T and above the vP phase, I propose a unified analysis of all the categorial types of ST in Jordanian Arabic, where ST is reduced to remnant movement of a constituent (NP, VP, AP, etc.), from which a subconstituent has undergone focus movement to an IP-internal focus position. The analysis captures all cases of ST in Jordanian Arabic, such as splitting with ditransitives. It accounts for the fact that ST in JA is only grammatical with bridge-contour intonation: the intonation is read off the syntactic structure in which functional projections such as TopP and FocP are encoded. The analysis provides insights into the structure of the left periphery of the vP phase: it lends support to the view that this area could have a more elaborate structure than previously thought, parallel to that of the left periphery of the CP phase.
1. Introduction
Topicalization is a syntactic operation in which a syntactic object (typically) representing given (old) information is internally merged in a sentence-initial position (1). Almost all natural languages exhibit topicalization of some form.
Some natural languages also display a special form of topicalization, in which, descriptively speaking, a constituent appears to be split into two parts and only one of them is topicalized; hence Split Topicalization (ST; e.g. Cerrone & Oda, Citation2019; Kahnemuyipour & Shabani, Citation2018). German is among the most studied languages in this respect:
In (2), bücher is topicalized, while other constituents, erst drei gute, that would otherwise form a constituent with it, remain in the base position. Following previous literature, I will refer to the topicalized constituent that appears sentence initially as TOP and the constituent that remains in the base position as REM.
Over the last four decades, different proposals have been made to account for ST, most of which mainly addressed data from German and Dutch, being the most studied languages for ST (e.g. subextraction, van Riemsdijk, Citation1989; distributed deletion, Fanselow & Ćavar, Citation2002; labeling-driven movement of an autonomous phrase, Ott, Citation2015). A major question that this work sought to answer is how TOP ended up in its position: does the derivation of ST involve movement of TOP (i.e. subextraction) or is ST derived by some other mechanism? Several empirical facts about ST are crucial in filtering out previous work. The first fact is that many cases of ST cannot have been derived via subextraction. One of these cases is what the literature refers to as gapless splits (3a), in which TOP does not associate with a gap in the constituent containing REM (3b).
Second, in contrast with the previous observation, other empirical evidence indicates that ST involves some form of movement. For instance, ST is sensitive to islands:
‘As for eyes, I know no woman who has more beautiful ones than I have.’
The third fact is that ST is characterized by special information structure properties and, generally speaking, is only acceptable with certain intonation patterns. In particular, TOP is characterized by a rising tone, and REM by a falling tone, which is referred to as bridge-contour splits (Ott, Citation2012). In this context, TOP is identified as a topic and REM as (contrastive) focus/new information.
The last fact is that ST is not restricted to nominals. Various categories can be split in topicalization. In addition to split noun phrase topicalization (SNPT; e.g. example (2)), languages display split verb phrase topicalization (SVPT; 5a) and split adjective phrase topicalization (SAPT; 5b), to mention just two.
Any analysis of ST should account for (at least) the first two facts which appear to be conflicting; ST does not seem to be derived by subextraction and at the same time it shows movement and connectivity effects. The analysis should also capture the special information structure reflected in the (often required) bridge-contour intonation. Ideally, the analysis should also capture all categorial types of ST. Previous work does not account for all of these facts. For instance, subextraction and distributed deletion analyses, which assume that TOP has formed a constituent with REM at some stage of the derivation prior to splitting, do not capture gapless splits. The labeling-driven analysis of Ott (Citation2015), which is based on the idea that TOP and REM have been merged together as autonomous phrases before TOP is eventually topicalized, perfectly capture the first two facts, but is designed for SNPT and does not address the information-structure properties of ST (but note that Ott builds on Ott, Citation2012 in which it is shown that not all cases of ST in German require a special intonation pattern).
This paper contributes to the work on ST by empirically and theoretically investigating ST in Jordanian Arabic (JA), a previously undiscovered phenomenon in this language. In (6), a nominal core is topicalized leaving a remnant in the base position, parallel to the cases of German and Dutch:
The paper reports on a large-scale acceptability judgment task of ST constructions of JA (n = 463; Section 2). The task reveals that JA displays ST of different categorial types. Further properties of ST in JA are also presented. Among these properties is the fact that ST in JA is characterized by bridge-contour intonation, with a rising tone falling on TOP and a falling accent on REM. To capture these facts, I propose a unified analysis of ST in which REM undergoes focus movement to the specifier of a low focus phrase in the IP domain (to the left periphery of vP; e.g. Belletti, Citation2004; Gengel, Citation2013; Jayaseelan, Citation2001; Rizzi & Bocci, Citation2017), which is followed by topicalization of the remnant of that movement. I illustrate the analysis for (6) in (7).
Table
I show that this analysis captures all the categorial types of ST in JA (see Section 3), including recalcitrant data, such as splitting with double-object constructions. The empirical investigation and the analysis proposed, if in the right direction, provide insights into the functional structure of the left periphery of the vP phase and forms a further aspect of the parallelism between CP and vP as phases; just like the left periphery of CP, the left periphery of vP could have a more articulate functional domain than previously thought.
2. The empirical picture of split topicalization in JA
This section presents the empirical facts of ST in JA, based on a large-scale acceptability judgment task.
2.1. Acceptability judgment task
As a result of the growing interest in collecting linguistic judgments in a scientific way through systematic linguistic experiments, instead of relying solely on informal judgment collection (cf. Al Khalaf & Mashaqba, Citation2023; Al Khalaf et al., Citation2024; Lau et al., Citation2017; Marty et al., Citation2020; Mashaqba et al., Citation2023; cf. Sprouse & Almeida, Citation2017), and for the study to be based on a rigid empirical basis, I decided to collect linguistic judgments in a systematic way for all the constructions in the paper. To that end, I designed an experiment to investigate the status of split topicalization in JA. The experiment is an acceptability judgment task that consists of a total of 94 sentences divided into four lists, each of which was mixed with fourteen fillers that are not related to split topicalization to ensure that the participants are attending to the task while doing the experiment. Only half of the fillers in each list were grammatical. The fillers were simple sentences that any speaker of JA can easily identify as grammatical/ungrammatical. To avoid any effect of the order of presentation of test items on the judgment, the items were randomized for each participant. Each list thus made an independent survey that was distributed to the participants online via Google Forms. I should note that the experiment does not have a factorial design; that is, it is a simple large-scale acceptability judgment task.
As for the participants, 463 native speakers of JA were recruited (individually, over a period of three months) to participate in the study. Most of the participants were undergraduate students. The task of the participant was to rate sentences on a 7-point Likert scale, with 1 corresponding to a very unacceptable sentence, and 7 to a perfectly grammatical one. Each participant individually completed one of four online surveys. Participants were excluded from the analysis if they scored less than 75% on the filler items; 40 participants were excluded for this reason. They were also excluded if they systematically gave a low rating (i.e. less than 3, assuming that 3.5 is the cutoff point of grammaticality) to all the sentences involving topicalization; thus, their results would not help in making generalizations about split topicalization; 27 speakers were excluded for this reason. This left 396 in all of the four surveys combined (divided as follows: 114, 99, 80 & 73). Of those, 73 were male and 323 were female. A large number of speakers were from the capital of Jordan, Amman, and neighboring areas, like Al-Balqaa and Az-Zarqaa. A good number was from Irbid, a northern area.
As the study involved an acceptability judgment rating task, no sophisticated statistical analysis was required. All that was required is the average rating of each sentence, which I show below to the right of each sentence between parentheses.
2.2. Categorial types of ST in JA
As is the case in the languages that display split topicalization, such as German and Dutch (Cerrone & Oda, Citation2019; Kahnemuyipour & Shabani, Citation2018; Ott, Citation2015; van Riemsdijk, Citation1989; van Hoof, Citation2006), the empirical investigation of JA reveals that various categories may be split in topicalization (many of the examples below involve sentential negation; see Al-Shawashreh et al., Citation2024 for a recent perspective on sentential negation in JA.) The most common categorial type is SNPT. Below are some examples:
Speakers seem to find SNPT more acceptable if TOP is definite (cf. 11b). This can be understood if we assume, following Fanselow and Ćavar (Citation2002) and van Hoof (Citation2006), that split topicalization consists of certain information structure properties (but see Ott, Citation2012 for an alternative view; cf. Kahnemuyipour & Shabani, Citation2022), such that REM has to be focused and TOP has to receive a secondary focus. Predicates also display ST; the first form is AP splits, where A occupies a topic position, while its complement/adjunct appears in the base position of the AP:
Note that in (13a) and (13b), the adjective is followed by an auxiliary verb (a kaan-class verb). I assume that this verb occupies a functional position above TP (e.g. Soltan, Citation2006).
Besides, JA displays split VP topicalization (SVPT) of the German type. In SVPT in JA, a core part of a VP appears sentence initially, and at least one verbal dependent appears in the base position of the VP, as illustrated in the b-examples below:
One might argue that the fronted verb has undergone head movement to some functional projection, given that in JA main verbs may undergo head movement. Though this can be said about some of the cases of SVPT, it does not seem to be right for many others. In (16b), V plus an indirect object can be topicalized. The same can be said about the examples in (17) and (18). If SVPT were just a case of head movement, these constructions should not be allowed. Thus, a head movement analysis does not seem to be the right analysis for SVPT.
Thus, as shown above, JA displays ST of different categorial types: SNPT, SAPT and SVPT.
One might wonder whether SNPT, SVPT and SAPT exhibit the same phenomenon. Previous literature distinguishes between topicalization and split-topicalization using different diagnostics. While the literature on SNPT focused on syntactic as well as semantic diagnostics, the literature that discussed the other categories depended on syntactic diagnostics. In particular, SNPT is characterized by a special semantic relation between TOP and REM; as shown in previous literature, TOP can have a subset-superset or part-whole relation with REM. At the same time, syntactically speaking, in split-topicalization ‘a constituent’s core is extracted to the prefield while leaving its non-core behind’ (van Hoof, Citation2006, pp. 441). More specifically, a core part of a complex (extended) projection is extracted, leaving the rest of the (functional) extended projection behind. Obviously, this is different from (pure) extraction that we see in topicalization.
Thus, as far as the differences between topicalization and split topicalization are concerned, the facts discussed in this section indicate that we are dealing with real cases of splitting.
2.3. Against subextraction
In this section, I show that ST in JA is not a case of subextraction. Subextraction is the earliest analysis of ST, in which ST results from the movement of TOP out of a projection that contains both it and REM (e.g. van Riemsdijk, Citation1989). This is based on the fact that many cases of ST constructions have corresponding non-split versions, where TOP could form a continuous constituent with REM (what Ott, Citation2012 refers to as the simplest cases).
Probably the strongest evidence against a subextraction analysis is when TOP appears to be entirely independent of REM (e.g. Ott, Citation2015; van Hoof, Citation2006). An example is gapless split topicalization, where it appears to be impossible for TOP to have formed a constituent with REM at any stage of the derivation:
Additionally, as indicated by Ott (Citation2012), many cases of ST violate typical constraints on extraction. This can be immediately seen in many cases of SNPT (e.g. 8b &9b); a nominal can be split in topicalization from its postnominal modifier although a nominal cannot undergo wh-movement, stranding the modifier:
Furthermore, certain mismatches in ST are hard to explain under a subextraction analysis. To take an example, the analysis does not account for number mismatches, where TOP is plural, while REM is an incompatible singular quantifier (and consequently cannot have formed a constituent with TOP), as in (24). If cases like this were derived via subextraction, non-split versions should be acceptable, contrary to fact (25).
Another issue that subextraction does not readily explain is the order preservation in case of topicalization of a string of more than one word (e.g. van Riemsdijk, Citation1989; van Hoof, Citation2006), such as when a V is fronted with a dependent with which it does not form a constituent (see Section 3.2).
Therefore, given the facts discussed above, a subextraction analysis of ST is implausible, a finding that previous literature has arrived at for other languages (e.g. German; van Hoof, Citation2006; Ott, Citation2015). An alternative to this analysis could be to posit full movement of the constituent in which splitting applies, followed by distributed deletion at PF in the higher and the lower copies (e.g. Fanselow & Ćavar, Citation2002). While this analysis explains why some cases of ST circumvent usual constraints on extraction, it does not capture the gapless splits or the agreement mismatches discussed above: it overgenerates and predicts that these cases are derived from ungrammatical sources in which TOP and REM form a constituent.
2.4. Movement effects of ST
As shown above, a subextraction analysis is implausible and does not capture all the cases of ST in JA. However, a number of facts suggest that TOP and REM are in a movement dependency (I use the same data points discussed in Ott, Citation2015). First, although split topicalization in JA is unbounded, and thus TOP and REM can be separated by clause boundaries (26b), the distribution of TOP seems to be restricted by locality; the acceptability of the sentence decreases noticeably when REM appears inside a movement island, as illustrated by (27b), (28b), and (29b). This sensitivity to islands indicates that movement is involved in the derivation of ST.
Previous literature (e.g. van Hoof, Citation2006) on split-topicalization shows that in SVPT, there is an A-Bar dependency between TOP and REM and that this dependency shows connectivity effects like island effects though VP topicalization is not known to show such an effect, as shown in (31).
The empirical investigation also reveals that SVPT in JA is also sensitive to islands, as illustrated below:
Comparing the ratings of SNPT and SNPT out of islands, we can see that speakers find SVPT out of islands less acceptable than SNPT. It is not quite clear why, but it could be that the difference might not be due to islands per se. It could be due to the general preference for what is called in classical grammar ‘nominal sentences’; that is, there is a preference for sentences that begin with nouns over sentences that begin with verbs for most speakers, at least in JA.
Second, ST shows connectivity effects. For instance, ST in JA shows binding connectivity effects, as illustrated in (35b) for Binding Condition C.
In addition, semantic scope provides evidence for there being a movement dependency between TOP and REM. More specifically, the semantic scope of TOP is calculated depending on its reconstructed position (in the study, the test items testing for semantic scope were supplemented with a context that biases the interpretation to a particular reading). In (36b), for instance, a wide-scope reading of the universal quantifier kull ‘every’ over the existential quantifier (the possessive pronominal clitic uh ‘his’ within TOP) is available (though not the most preferred reading given the relatively lower average rating due to speaker variation in allowing this reading). This is only possible if TOP reconstructs to its base position (in the c-command domain of the universal quantifier).
Again, this indicates that there is a movement dependency between TOP and REM.
2.5. Interim summary
I drew the empirical picture of ST constructions in JA by reporting on the results of a large-scale acceptability judgment task. The results show that JA exhibits ST of various categorial types and that although ST cannot be derived via subextraction, the construction shows movement and connectivity effects.
3. Analysis
In this section, I present my analysis of ST constructions in JA. The analysis is based on the idea that there exists an IP-internal focus position, where REM moves to the specifier of this phrase. More specifically, following Jayaseelan (Citation2001), Belletti (Citation2004), Gengel (Citation2013), among others, there is an IP-internal FocP position, below T and above vP. Belletti (Citation2004) argues convincingly that new information focus occupies this position in languages like Italian. Jayaseelan (Citation2001) and Gengel (Citation2013) argue that this position hosts the elements that survive ellipsis in pseudogapping constructions. I follow these lines of research and propose that this position hosts REM in ST constructions. Similarly, previous literature showed that JA displays low focus position. For instance, in (37), where the construction is a CP, kaan occupies C position (e.g. see Soltan, Citation2006)).
Here ‘articles’ is a focused object of ‘publish’. That kaan occupies C necessitates that ‘articles’ is merged as focus IP-internally.
What confirms the focus movement of REM is that in JA, TOP is characterized by a rising tone, and REM by a falling tone, which Ott (Citation2012) refers to as bridge-contour splits (but see Ott, Citation2012 for different findings for ST in German). In this context, TOP is identified as a topic and REM as a (contrastive) focus/new information. Two of the surveys in the acceptability judgment task consisted of audio recordings of all of the constructions along with their written forms. Within these surveys, a number of the sentences involving ST were read with a bridge-contour intonation (e.g. 38a, 39a, and 40a) and with a flat intonation (e.g. 38b, 39b and 40b) to test for the effect of intonation on sentence acceptability (the recordings of each of these sentences were included in different surveys to make sure that the participant rates only one sentence of the minimal pair).
As can be seen, ST with flat intonation was given lower ratings, which indicates that the intonational pattern of the ST is crucial to grammaticality. Following Rizzi (Citation1997) and Belletti (Citation2004), I propose that these intonation patterns are read off the syntactic positions of constituents; bridge-contour intonation indicates that TOP occupies a topic position and REM a focus position. I show below how this assumption can be used to explain ST in JA. In section 3.1, I address SNPT. In section 3.2, I turn to the rest of the categorial types.
3.1. SNPT
One line of research on SNPT in German and Dutch argues that TOP and REM are generated as independent phrases before TOP is eventually topicalized. The earliest version of this type of analysis was proposed by Fanselow (Citation1988), where TOP and REM are generated as independent phrases, and where REM is an elliptical NP involving a null pro licensed by ellipsis. Ott (Citation2012, Citation2015) builds partly on this analysis and the labeling framework of Chomsky (Citation2013) and proposes that ST is derived via an asymmetrizing, labeling-driven movement of TOP. TOP is merged with REM (an elliptical nominal) as a set of autonomous phrases, {XP, YP}, resulting in a locally unstable unlabelable symmetric set. This symmetry must be broken and the combination must be labeled because labeling is required by the Conceptual-Intentional (C-I) interface where only labeled constituents can be interpreted. The labeling algorithm scanning the combination fails to identify a unique Z such that Z is a head in the symmetric set of {XP, YP}. Thus, according to Ott (Citation2015), as a last resort, XP moves to asymmetrize the set and consequently allow it to be labeled, assuming that the set takes the label of the phrase that is properly contained in the set, i.e. YP (Ott, Citation2012, Citation2015 also extends this analysis to quantifier float in German; see Al Khalaf, Citation2019 for an application of this analysis on Standard Arabic).
Adopting Ott’s analysis, I propose that in SNPT constructions, TOP and REM are merged together as a symmetric set of phrases, where, semantically speaking, TOP is a predicate taking REM as an argument. I propose, however, that it is REM that moves out of the set rather than TOP. REM moves to the specifier of FocP in the IP domain, which renders the symmetric set labelable. Then, the remnant of this focus movement moves (via the edge of the vP phase) to spec-TopP in the CP domain. To illustrate, consider example (41), reproduced from above.
It should be noted that we can understand the semantic relationship between ‘green beans’, and ‘beans’ as a subset-superset relationship. The derivation of the sentence proceeds as follows. As illustrated in (42), the phrases ‘the beans’ and ‘the green’ merge as a symmetric set. The derivation continues bottom-up until Foc is merged. REM (i.e. ‘the green’) then internally merges into the specifier of FocP (but note that the movement occurs successive-cyclically; thus a copy of REM should also be internally merged at the vP phase edge. I cut this step out for space limitations, however). This results in ‘the beans’ being the only phrase properly contained inside the symmetric set; thus, the labeling algorithm will identify this phrase as the label of the combination. Upon the merger of Top head, this remnant phrase is internally merged to spec-TopP. Note that the subject also merges as a specifier of a lower TopP (e.g. Soltan, Citation2006), assuming that TopP may iterate (Belletti, Citation2004; Rizzi, Citation1997). I assume that NegP is headed by the negative marker (e.g. Zanuttini, Citation1997, Citation2001), here ma, and scopes over TP (e.g. Rizzi, Citation1997, Citation2004).
Table
Note that the internal merger of TOP and REM to the left edges of TopP and FocP, respectively, applies freely, without the need for triggering features (and thus there would not be a violation of the Inclusiveness Condition; cf. Al Khalaf, Citation2024, Citation2022a, Citation2022b). Merge is free so long that the resulting constructions can be interpreted at the C-I interface (Chomsky, Citation2004, Citation2013).
The labeling-driven movement adapted here accounts for the movement and connectivity effects of ST. It also captures gapless and agreement mismatches in which TOP and REM are syntactically independent of each other. The often bridge-contour intonation required by JA speakers for ST to be acceptable is also explained as a result of the discourse-related positions that are syntactically encoded in the ST construction, namely TopP and FocP.
Thus, to recapitulate, SNPT is derived via two steps: (i) a focus movement, which results in labeling of the symmetric set containing TOP and REM, and (ii) topicalization of the remnant phrase of that movement. These two movements explain the correlation between ST in JA and the bridge-contour intonation.
3.2. SVPT and SAPT
I now turn to the other cases of ST, namely SVPT and SAPT. These differ from SNPT in that it is a predicate that appears to have moved away from at least one of its arguments. A labeling-driven movement, which was originally devised to account for SNPT, does not seem to extend to these cases. This is partly because that analysis is based on the assumption that REM is an elliptical NP, and is syntactically (but not semantically) independent of TOP. In SVPT and SAPT constructions, on the other hand, REM is not syntactically independent of TOP (as it is an internal argument of it). Thus, in SVPT, one would not want to say that the object and the verb merge as a symmetric set of independent phrases. The opposite is true; a set of V and its argument should always form an asymmetric set which should be labeled as VP.
An initial analysis of these cases, then, could be that TOP ends up in its position via extraction. This, however, is unlikely if we consider splitting in double-object constructions (e.g. 16b), where the verb plus the indirect object is TOP and the direct object is REM. Assuming that the indirect object occupies spec-VP and the direct object is the internal argument (cf. Bruening, Citation2010), extracting V plus the indirect object would be impossible because they do not form a constituent:
Table
One might argue that it could be that V and its dependent have moved separately. Although this analysis would capture the data, it misses the generalization that in all the cases of SVPT in which V is fronted with its dependent, the linear order of the string in the base position is always preserved in the fronted string (the same can be said about many cases of SNPT as well; van Hoof, Citation2006). Thus, further assumptions would have to be made to account for the linearization of TOP, which would make the analysis less parsimonious than one that does without such assumptions. At the same time, although subextraction is unlikely, the fact that SVPT is sensitive to islands as shown in section 2.4, indicates that TOP ended up in its position via movement (i.e. it is not base generated in this position, for instance).
I propose instead that SVPT is also derived via focus movement plus remnant movement as shown for SNPT above. Müller (Citation1998) proposes that SVPT in German is derived via remnant movement of VP, which is preceded by scrambling of REM. I follow the spirit of this analysis and propose that SVPT is derived via a similar mechanism, albeit with a focus movement of REM rather than scrambling. To illustrate my analysis, consider example (44), modified from above, which is a simple case of SVPT involving topicalization of V:
As illustrated in (45), the QP ‘everything’ undergoes focus movement via the vP phase edge of spec-FocP. Then the remnant VP, ‘buy’, is then topicalized (I suggest that kann moves to Top head via head movement; cf. Soltan, Citation2006).
Table
Now consider example (46), modified from above, in which splitting occurs in a double-object construction:
As indicated above, cases like this one are challenging because of two reasons. First, as mentioned earlier, V plus the indirect object are fronted, leaving the direct object behind, which cannot be explained under subextraction without further assumptions. Second, the verb is under v, while the indirect object is at the left edge of VP; thus, an analysis based on remnant VP topicalization does not capture this case. I suggest that cases like this provide evidence that the vP phase may undergo topicalization if the topicalized constituents fall outside the VP projection, but are still within the phase being built. I thus assume that the following generalization holds:
Given the above informal generalization, which I assume falls under the general umbrella of economy, the derivation of (46) proceeds as follows. After the VP phrase has been fully derived and upon the merger of v head, the verb is internally merged under v. The derivation proceeds in the same way explained above; the REM ‘many prizes’ undergoes focus movement. Then when Top is merged, the whole vP phase is internally merged to the specifier of TopP because the phase is the smallest projection that contains the elements to be topicalized (i.e. v plus the indirect object).
Table
The analysis also accounts for SAPT; REM undergoes movement and then the remnant AP is topicalized. Consider the following example, repeated from above:
As illustrated in (50), ‘with all his decisions’ moves out of AP to spec-FocP. The remnant AP, ‘satisfied’, moves to Spec-TopP.
Table
3.3. Interim summary
I proposed a unified analysis of all categorial types of ST in JA. ST occurs as a result of focus movement of REM to the specifier of a focus phrase in the low IP domain (in the left periphery of vP), followed by topicalization of the remnant phrase to the specifier of TopP in the CP domain. The analysis accounts for the peculiarities of SNPT by adopting a modified version of the labeling-driven movement of Ott (Citation2015), such that the focus movement in SNPT occurs out of a symmetric set of autonomous phrases. The analysis also explains why ST in JA is only acceptable with bridge-contour intonation: ST constructions have syntactically encoded information-structure constructs, namely FocP and TopP.
4. Conclusion
In this paper, I empirically and theoretically investigated split topicalization in Jordanian Arabic, a previously undiscovered phenomenon in this language. The results of the investigation reveal that different categorial types can be split in topicalization in Jordanian Arabic. They also show that although ST is restricted by locality, it cannot have been derived via subextraction. It was found that generally, ST is acceptable only with bridge-contour intonation.
To account for these facts, I proposed a unified analysis of all categorial types of ST by adopting the view that there exists an IP-internal focus phrase in the left periphery of the vP phase (e.g. Belletti, Citation2004; Gengel, Citation2013; Jayaseelan, Citation2001), the spirit of the remnant movement analysis of SVPT proposed by Müller (Citation1998), and the labeling-driven movement analysis of SNPT advanced in Ott (Citation2015). More specifically, I proposed that ST occurs as a result of two movements: focus movement of REM to spec-FocP in the IP domain, followed by topicalization of the remnant of that movement to spec-TopP in the CP domain; thus, all instances of ST are reduced to cases of remnant movement.
The analysis provides further support for positing a FocP in the left periphery of the vP phase. This, in turn, adds a further aspect of parallelism between CP and vP as phases and suggests that it is more likely that the vP edge has a more elaborate functional domain than previously thought. The analysis also adopts the assumptions of the cartographic program and includes them in a minimalist analysis, but keeps the fundamental assumption that merge should be free to apply, without the need for triggering features, abiding by the Inclusiveness Condition (cf. Bruening & Al Khalaf, Citation2020).
Disclosure statement
No potential conflict of interest was reported by the author(s).
Additional information
Notes on contributors
Eman Al Khalaf
Eman Al Khalaf is an associate professor of linguistics at the University of Jordan, specializing in experimental syntax and semantics. She earned her MA and Ph.D. degrees in linguistics and cognitive science from the University of Delaware.
References
- Al Khalaf, E. (2019). Floating quantifiers are autonomous phrases: A movement analysis. Glossa: A Journal of General Linguistics, 4(1), 1–23. https://doi.org/10.5334/gjgl.848
- Al Khalaf, E. (2022a). Furthest conjunct agreement in Jordanian Arabic: Evidence for multiple (non)simultaneous Agree. Natural Language & Linguistic Theory, 40, 345–391. https://doi.org/10.1007/s11049-021-09515-0
- Al Khalaf, E. (2022b). Constituency and left-sharing in coordination. English Language and Linguistics, 26(1), 161–183. https://doi.org/10.1017/S1360674321000162
- Al Khalaf, E. (2024). Semantic scope ambiguity in gapping and non-constituent coordination: A generative analysis. Cogent Arts & Humanities, 11(1), 1–9. https://doi.org/10.1080/23311983.2024.2322231
- Al Khalaf, E., & Mashaqba, B. (2023). Semantic scope in Q-float constructions: An experimental investigation. Lingua, 282(2023), 103457. https://doi.org/10.1016/j.lingua.2022.103457
- Al Khalaf, E., Mashaqba, B., & Aldiqs, R. (2024). The syntax of non-canonical coordination in Jordanian Arabic: An experimental investigation. Open Linguistics, 10(1), 1–14. https://doi.org/10.1515/opli-2024-0001
- Al-Shawashreh, E., Jarrah, M., & Al Khalaf, E. (2024). Variation in verbal negation in Jordanian Arabic: A corpus-based analysis. Lingua, 297, 103644. https://doi.org/10.1016/j.lingua.2023.103644
- Belletti, A. (2004). Aspects of the low IP area. In L. Rizzi (Ed.), The structure of CP and IP: The cartography of syntactic structures (Vol. 2, pp. 16–51). Oxford University Press.
- Bruening, B. (2010). Double object constructions disguised as prepositional datives. Linguistic Inquiry, 41(2), 287–305. https://doi.org/10.1162/ling.2010.41.2.287
- Bruening, B., & Al Khalaf, E. (2020). Category mismatches in coordination revisited. Linguistic Inquiry, 51(1), 1–36. https://doi.org/10.1162/ling_a_00336
- Cerrone, P., & Oda, H. (2019). Reinterpreting ne-cliticization as split-topicalization. University of Pennsylvania Working Papers in Linguistics, 25(1), 51–59. https://repository.upenn.edu/pwpl/vol25/iss1/7
- Chomsky, N. (2004). Beyond explanatory adequacy. In A. Belletti (Ed.), Structures and beyond: The cartography of syntactic structure (Vol. 3, pp. 104–131). Oxford University Press.
- Chomsky, N. (2013). Problems of projection. Lingua, 130, 33–49. https://doi.org/10.1016/j.lingua.2012.12.003
- Fanselow, G. (1988). Ufspaltung von NPn und das problem der ‘freien’ wortstellung. Linguistische Berichte, 114, 91–113.
- Fanselow, G. (1993). The return of the base generators. Groninger Arbeiten Zur Germanistischen Linguistik, 36, 1–74.
- Fanselow, G., & Ćavar, D. (2002). Distributed deletion. In A. Alexiadou (Ed.), Theoretical approaches (pp. 65–109). John Benjamins.
- Gengel, K. (2013). Pseudogapping and ellipsis. Oxford University Press.
- Jayaseelan, K. A. (2001). IP-internal topic and focus phrases. Studia Linguistica, 55(1), 39–75. https://doi.org/10.1111/1467-9582.00074
- Kahnemuyipour, A., & Shabani, M. (2018). Split noun phrase topicalization in Eshkevarat Gilaki. The Linguistic Review, 35(4), 625–646. https://doi.org/10.1515/tlr-2018-2003
- Kahnemuyipour, A., & Shabani, M. (2022). Corrigendum to: Split noun phrase topicalization in Eshkevarat Gilaki. The Linguistic Review, 39(2), 333–333. https://doi.org/10.1515/tlr-2022-2089
- Lau, J. H., Clark, A., & Lappin, S. (2017). Grammaticality, acceptability, and probability: A probabilistic view of linguistic knowledge. Cognitive Science, 41(5), 1202–1241. https://doi.org/10.1111/cogs.12414
- Marty, P., Chemla, E., & Sprouse, J. (2020). The effect of three basic task features on the sensitivity of acceptability judgment tasks. Glossa: A Journal of General Linguistics, 5(1), 72. https://doi.org/10.5334/gjgl.980
- Mashaqba, B., Al Khalaf, E., Huneety, A., & Abu Sa’aleek, H. (2023). Subject-verb agreement inflection in Arabic-speaking individuals with Down syndrome. Clinical Linguistics & Phonetics, 1–22. https://doi.org/10.1080/02699206.2023.2221373
- Müller, G. (1998). Incomplete category fronting: A derivational approach to remnant movement in German. Kluwer.
- Ott, D. (2012). Local instability: Split topicalization and quantifier float in German. de Gruyter.
- Ott, D. (2015). Symmetric merge and local instability: Evidence from split topics. Syntax, 18(2), 157–200. https://doi.org/10.1111/synt.12027
- Rizzi, L. (1997). The fine structure of the left periphery. In L. Haegeman (Ed.), Elements of grammar (pp. 281–337). Kluwer.
- Rizzi, L. (2004). Locality and left periphery. In A. Belletti (Ed.), The cartography of syntactic structures, vol. 3, structures and beyond (pp. 223–251). Oxford University Press.
- Rizzi, L., & Bocci, G. (2017). Left periphery of the clause: Primarily illustrated for Italian. In M. Everaert & H. C. Riemsdijk (Eds.), The Wiley Blackwell companion to syntax, second edition. John Wiley & Sons. https://doi.org/10.1002/9781118358733.wbsyncom104
- Soltan, U. (2006). Standard Arabic subject-verb agreement asymmetry revisited in an agree-based minimalist syntax. In C. Boeckx (Ed.), Agreement systems (pp. 239–265). John Benjamins.
- Sprouse, J., & Almeida, D. (2017). Design sensitivity and statistical power in acceptability judgment experiments. Glossa: A Journal of General Linguistics, 2(1), 1–32. https://doi.org/10.5334/gjgl.236
- van Hoof, H. (2006). Split topicalization. In M. Everaert & H. C. van Riemsdijk (Eds.), The Wiley Blackwell companion to syntax (pp. 408–462). Blackwell.
- van Riemsdijk, H. (1989). Movement and regeneration. In P. Beninc a (Ed.), Dialect variation and the theory of grammar. Foris.
- Zanuttini, R. (1997). Negation and clausal structure: A comparative study of Romance languages. Oxford University Press.
- Zanuttini, R. (2001). Sentential negation. In M. Baltin & C. Collins (Eds.), The handbook of contemporary syntactic theory (pp. 511–535). Blackwell.
![Supercharge Your Next Research Paper: Research and write your next paper with Jenni AI.]({"type":"image" , "src" : "/sda/112441/MPU-humanities-jenniai.png"})