Some Initial Remarks on Non-Prepositional Genitives in the Apulian Variety of San Marco in Lamis

: This work aims at an initial description of prepositionless geni-tives in the Romance variety of San Marco in Lamis, spoken in the Southern Italian region of Apulia. The construction will be compared with other Romance, Semitic, Albanian, and Iranian varieties whereby the expression of possession is connected to the presence of D elements, or to morphology stemming from them. The paper deals, in particular, with the behaviour of the construction with elements such as definite and indefinite articles, demonstratives, proper names, and with how pre-nominal adjectival modification of geni-tives and post-nominal adjectival modification of heads can only occur in the prepositional kind of the construction. This is also the case with demonstratives preceding heads and genitives in the form raised nominals. It will be seen that genitives are only interpreted as such when they are non-raised, i.e. when they are articled. The pre-genitival article is thus understood to be a pivotal element in the interpretation of the second DP as genitival.


The data 1.1 Core elements: prepositions, determiners, and proper names
One of the first examples of a non-prepositional genitive in the dialects of Southern Italy quoted in the literature was Calabrian, going back to Rohlfs (1966Rohlfs ( [1949]]), which describes the phenomenon in the Calabrian dialect of Morano ascribing it to preposition absorption.More recent studies on Calabrian varieties like Silvestri (2012;2016) challenge instead an account on the lack of the preposition on phonetic grounds such as the one Rohlfs puts forward.Simply put, a non-prepositional genitive is a genitival construction whose argument is not introduced by a preposition.Generally speaking, since we are moving within the realm of Romance languages, genitive prepositions are elements deriving from Latin de resulting in various forms.For instance, in Neapolitan this resulted in the preposition e when in its isolated form.If combined with definite articles, it yields r-a ('of-the.F.SG.'); r-e ('of.the.F.PL.'); r-o ('of.the.M.SG.'); r-i ('of.the.M.PL.')where the vocalic exponent corresponds to the definite article, while r-(>d-) corresponds to the prepositional element.In those cases where such prepositions do not introduce the possessor we speak of non-prepositional genitives.In the variety analyzed here de took the form of də ('of') when isolated; combined with definite articles it yielded də-lla ('of.the.F.SG.'); də-llu ('of.the.M.SG.'), də-lli ('of.the.PL.').
As for the present variety, the first and probably most evident constraint concerns the definiteness features of both the head and the genitive DP: (1) a. la koda lu kanə the tail the dog 'the tail of the dog' b. *la koda nu kanə the tail a dog '*the tail of a dog' c. la koda də nu kanə the tail of a dog 'the tail of a dog' d. *na rɔta la makəna a wheel the car '*a wheel of the car' e. na rɔta dəlla makəna a wheel of.thecar 'a wheel of the car' It is easy to spot a resemblance with the definiteness feature requirement of Semitic Construct State nominals (Borer 1988;Dobrovie-Sorin 2000;Danon 2007;Borer 2012;Shlonsky 2012, among others) and especially its Hebrew type, except for the fact that in this Romance variety, the head of the genitival complex is not articleless -with the exception of those instances in which the head is in the vocative case.
(2) ah, servə li padrunə oh servants the masters 'oh, servants of the masters' As the Southern Italian under consideration does with də, Hebrew resorts to the lexicalization of the preposition šel in the case of indefinite Ns (examples from Danon 2008: 902): 1 1 The presence of šel does not imply that Ns are interpreted as indefinite (Danon 2007, where the following data are taken from): (i) a.
ha-xulca šel ha-yeled nirteva the-shirt of the-boy got.wet 'the boy's shirt got wet' As a general rule, though, we can say that the head noun inherits the definiteness features of the genitive noun, which led to the stipulation of a mechanism involving upwards percolation (on this particular point see Borer 1988;Shlonsky 2012) to account for the phenomenon of definiteness spreading.On the other hand, the indefiniteness features of the genitive noun do not seem to spread to the head noun (Borer 1988).
(3) a. tmunat ha-yeled picture the-boy 'the picture of the boy' b. tmuna šel ha-yeled picture of the-boy 'a picture of the boy' Furthermore, in the variety of San Marco in Lamis, proper names as genitive arguments are allowed only when introduced by a preposition.One of the hypotheses being made in the literature on proper names is that they rise from N to D (Longobardi 1994).The obligatory lexicalization of the preposition in front of arguments which are proper names is directly correlated to the lack of an overt definite article.The fact that the preposition is needed where a definite article is not present suggests that the definite article preceding the genitive might be a pivotal element in the interpretation of the latter as a possessor.
An exception to (4) involves a group of proper names which do not seem to undergo N to D raising, i.e. la partita la Juventus 'Juventus match', lu koncertə li Metallica 'Metallica's concert'.Apart from being an exception when it comes to the lexicalization of the prepositional element, these examples show that the construction is quite productive, and used by young speakers as well, as the nouns entering in the construction show, which include modern referents, i.e.: la crack lu jokə 'the code to crack the videogame'.With regards to other Dialects of Southern Italy, Silvestri (2016) reports instead on prepositionless genitives in the Calabrian variety of Verbicaro, which are perceived as archaic.
Going back to the Apulian construction, a first feature allowing its realization is, as previously said, definiteness.Yet definiteness alone is not enough.It is clear that it is a combination of features we are dealing with.Consider the following data from the type of San Marco in Lamis: (5) a. la rɔta la makəna the wheel the car 'the wheel of the car' b. *na rɔta la makəna a wheel the car 'a wheel of the car' c. na rɔta dəlla makəna a wheel of.thecar 'a wheel of the car' d. l oɲɲa li pedə the nails the feet 'the feet's toenails' e. li tsuntsə lu balləkettə the icicles the balcony 'the icicles on (of ) the balcony' First, both the head and the argument noun can be singular and plural.So, no constraint is placed on number features.Indefinite nouns, as stated, are not allowed in the construction.Definite nouns, on the other hand, are allowed, provided that the determiner they are preceded by is not a demonstrative, neither a distal (6b), nor a proximal one (6d):

Container vs. part-whole
Consider now the data in (8).In (8a) the preposition may not appear.However, the same nouns that in (8a) appear within a non-prepositional genitival complex, in (8b, d) must occur with a preposition.This distinction shows that the two examples must correspond to different types of genitival modification.A cup of coffee may refer either to a quantity of drinkable coffee inside the cup, or to the cup that contains it.In order to distinguish between the two, we embedded the genitival complex in the argumental slot of verbs that take contained or container as their object.Speakers judged as ungrammatical sentences with clean when the genitival complex did not include a preposition (8b, d), whereas drink yields a different pattern( 8a The preposition can occur in any type of genitival complex.It is its lack, that is constrained.For the time being, however, I will limit myself to simply describing such a distinction and leave the contrast in ( 8) as an open question.

Modification, head-argument adjacency, coordination of the genitive noun, and constituent isolation
When it comes to adjectival modifiers, with the exception of a handful of cases like bonə/bɔna ('good.m.sg.'/'good.f.sg.'), bellə/bɛlla ('beautiful.m.s g.'/'beautiful.f.sg.'), and bruttə/brutta ('ugly.m.sg.'/'ugly.f.sg.'), which can be pre-nominal, as it is common through the Dialects of Southern Italy, the rest obeys a strict post-nominal restriction, including adjectives expressing size (rossə/rɔssa-'big', and məninnə/mənenna -'small'; Standard Contemporary Italian, instead, also includes such adjectives among those allowed to occur in pre-nominal position).This premise on the positioning of adjectives is necessary in order to understand some core features of the construction considered here.In particular, the main point concerns modification of the head.
In prepositionless genitives, post-nominal adjectival modification of the head noun often triggers the realization of a second type of genitival construction.Consider Semitic, specifically Hebrew again, for example, with data from Dobrovie-Sorin (2000: 195): (9) a. beyt ha-iš house the-man 'the man's house' b. ha-bayit ha-gadol *(šel) ha-iš the-house the-big *(of) the-man 'the big man's house' Hebrew Construct State nominals introduce the possessor with a definite article but also require the head and its argument to be adjacent.Once the head is followed by a modifying adjective, the argument must be preceded by a preposition (šel).
In the Apulian variety of San Marco in Lamis, modifiers interposed between the head and the argument of prepositionless genitives yield the realization of the preposition də, as in (10a).Post-nominal adjectives might only follow the argument (10c): (10) a. lu libbrə novə dəllu prəssorə the book new of.theprofessor 'the new book of (assigned by) the professor' b. *lu libbrə novə lu prəssorə the book new the professor '*the new book of (assigned by) the professor' c. lu libbrə lu prəssorə novə the book the professor new 'the book of (assigned by) the new professor' Alternatively, the head can undergo adjectival modification, provided that the adjective modifying the head belongs to the set of pre-nominal ones, so as not to break the adjacency between the head and the argument.Witness ( 11): Heads in Italian N+N genitival compounds like casa Rossi also tend to resist modifiers interposed between the head and the argument (12a), like Construct State nominals.This was noted in Longobardi (2001) which shows that the adjective can modify the head, but only when occurring on the right of the genitival complex (Longobardi 2001: 572) 12), in the variety of San Marco in Lamis adjectives might appear to the right of the genitival complex, but unlike in Italian (12b), they do not modify the head, but rather the genitive.This is exemplified in ( 13) by gender features appearing on the modifying adjective.On the other hand, adjectival modifiers appearing after the genitive noun of Hebrew Construct State nominals might be understood as modifying either the head noun or the genitive, as reported in Borer (1988).Adjacency requirements between the head and the genitive noun led to a proposal put forward in Siloni ( 2003) where genitive case is checked at PF in prosodic terms.Modification is also what determines the choice of the genitival construction to be employed in Romanian.In this language two kinds of genitives might be distinguished.The first involves an N-D N-D string, with genitive case marked on the enclitic definite article of the genitive, e.g.vecinului in ( 14).The second involves the al series of linker elements (a, in the feminine singular) preceding the genitive, cf. ( 14 Dobrovie-Sorin (2000) considers genitives to occupy the SpecDP position -which is rightwards-oriented.Thus in ( 16), fata vecin-ului, 'the neighbor's daughter' the head noun, fata, is thought by Dobrovie-Sorin to be hosted by D alongside the definite article, with the genitive occupying the SpecDP position.Genitive case is assigned only if N carries a definite article (and not a demonstrative), so it follows that the definite article is necessary, in her view, to assign case.The adjunction is justified by Dobrovie Sorin also on the grounds that al genitives can appear in isolation.What she calls purely synthetic (al-less) genitives, on the other hand, cannot (see also Cornilescu 1995: 18) Non-prepositional genitival arguments are only allowed when they are pronounced along with their head.Failure to do so will trigger a non-oblique interpretation of the DP, either as nominative or accusative.
Prepositionless genitives cannot undergo coordination.The coordinated possessor must thus be preceded by a preposition (20b), suggesting once more the strict adjacency requirements of the construction: (20) a. *l oɲɲa li manə e li pedə the nails the hands and the feet ' *the nails of the fingers and the toes' b. l oɲɲa li manə e dəlli pedə the nails the hand and of.thefeet 'the nails of the fingers and of the toes'

Lack of P in non-genitival constructions: comparatives and Qualitative
Binominal Noun Phrases In the preceding section, we have seen that genitives not introduced by a preposition may never occur in isolation.In fact, this turns out to be possible in one case only, which concerns possessives.In the variety of San Marco in Lamis, possessives are always articled, when occurring with the noun (21, a), after a copular expression (21 a, b, c), and when in isolation ( 21, d As we said earlier this variety does have the possibility of using a preposition.The example in ( 22) can also be uttered with the preposition də.The lack of the preposition in ( 22) is as not constrained as in the cases of prepositionless genitives reviewed so far, but it is relatively free and occurs with possessives (22a) and articled nouns (22b) alike.Clearly this is a different phenomenon from possessors.However, we also find that a lack of the preposition occurs in a number of non-genitival contexts where the preposition də can also be used.In particular, in comparison constructions where a noun can be compared against a set comprising one (22a) or several elements (22b) through a preposition relating the two.
Constructions such as Qualitative Binominal Noun Phrases (Den Dikken 2006; Kayne 1994) are yet another instance in which the preposition might not occur.Den Dikken (2006) proposes an account of Qualitative Binominal Noun Phrases in which they are derived from a copular construction whereby the predicate inverts with its subject.Thus, a sentence like a jewel of a village is thought to be deriving from the village is a jewel.The predicate, jewel, inverts its position with the subject, village.The preposition preceding the subject in the final linear order is then realized as a syntactic aid to the inversion of the predicate.In this sense, the preposition of is seen as a nominal copula.In the variety of San Marco in Lamis, binominal noun phrases occur as in ( 23 Clearly what lacks in (23 a-b) is the preposition, the very element that in Den Dikken's predicate inversion is realized as an inversion-aiding 'device'.Den Dikken's proposal cannot be applied to (23), given the absence of the preposition, unless the inversion-aiding preposition is characterized as an empty element.

On determiners and genitival modification
Here we shall recapitulate some relevant findings in linguistic research on possession and its correlation with determiners, including both demonstratives and definite articles.
Starting with Afroasiatic, Pennacchietti (1968) shows how genitive morphological marking in Semitic is ultimately ascribable to ð/z determiners.In Neo-Syriac possession is in fact expressed with the genitive enclitic marking -t on the possessee and proclitic marking on the possessor (Pennacchietti 1968: 32): He also notes how in the Jewish Neo-Aramaic of Persian Azerbaijan i-d also came to introduce restrictive relatives, ktab-i-d b-idew, 'the book that (was) in his hand.'The same applies to Modern Hebrew, where the preposition šel ('of') contains the same D base, which also gave rise to relative pronouns, and possessives (Pennacchietti 1968: 10-11).This pattern is not restricted to Afroasiatic but is also present in the Dravidian family, with Telegu and Classical Tamil genitive case morphology being linked to D elements (Caldwell 2013(Caldwell [1856]]).
In Balkan languages such as Aromanian and Albanian (and in varieties of the latter) possessive constructions make use of an element known as linker (Den Dikken and Singhapreecha 2004, Manzini et al. 2014, Franco et al. 2015, among others) in linguistic literature.The morphological make up of linker elements often includes a D base, coincides with it, or it is an allomorph of it.
Here we can see an example from the Arbëresh variety of Vena di Maida, spoken in the Italian region of Calabria, with data from Franco et al. (2015: 280): (25) biʃt-i i matʃɛ-sə tail-NOM.M.def LKR.M cat-OBL.F.def 'the tail of the cat' The linker i in (25) agrees in gender with the head noun biʃt-, 'tail', and this can be seen from definite morphology -i attached to 'tail', identical to the linker, which is therefore a D element, 3  Unlike the Albanian linker in (25), the Aromanian linker (26) agrees with the genitive, rather than with the head, and does so in case and φ-features.
Indo-Iranian linkers, in languages such as Farsi and Kurdish, agree with the head noun when morphological agreement is present (as in Kurdish).Still in Kurdish (Kurmanji) the ezafe morpheme which is employed in genitival contexts can also have an anaphoric function, akin to that of a determiner (in Meillet 1931 it is also hypothesized that the ezafe morpheme might be traced back to hya, the Old Persian demonstrative pronoun).This is exemplified in ( 27) for Kurmanji Kurdish (Mackenzie 1961: 163;Manzini et al. 2014: 240) On the basis of such data (and data on patterning with Romance clitic pronouns), in the latter work the Iranian ezafe too is considered to be a D element and part of a construction denoting inclusion of the possessee within the possessor in a part-whole type of relation.
Let us now consider a couple of examples of the Chinese linker de from Simpson (2002: 12) (28) a. wo zuotian mai de nei-ben shu I yesterday buy LKR that-CL book 'the book I bought yesterday' b. wo de nei-ben-shu I LKR that-CL-book 'that book of mine' As (28) shows, de appears after a pre-nominal relative clause (28a) and after a possessor (28b).Basing himself on cross-linguistic evidence on modifying phrases appearing with their own determiners (such as Balkan and Hebrew pre-genitival and adjectival linkers), Simpson suggests that such a linker might actually be a definite determiner.More precisely, he states that the linker is actually an element that is similar to the Romanian enclitic article, in that in possessive constructions the possessor is in a pre-de position (the same happens for relative clauses), the same position in which Romanian lexicalizes the noun with respect to its definite article.He then proposes the following structure for de constructions (ibidem): ( 29 Diachronically, de can be traced back to the Classical Chinese determiner zhi, which Simpson describes as an element having a parallel distribution to that of Modern Chinese de, except that it could also be used as a demonstrative (Modern Chinese de is reported instead as an element having lost such deictic function), as in (30) from Simpson (2002: 17) Simpson thus suggests that the nature of determiners themselves might allow them to instantiate syntactic variables permitting a number of elements (whether adjectival or genitival, but this also extends to relative clauses) to enter in a modifying relation with a nominal.The hypothesis is also based on crosslinguistic data showing that a further determiner is often needed only in cases where a modifying element appears.This includes, as we said, adjectival and genitival linker constructions of the types previously discussed, but also Hungarian data he considers from Szabolcsi (1994) in (30) and English ones in (31) with relative clauses (30, b) and possessor phrases (30 c-d, 31 a-b) (Simpson 2002: 15, 19 The observation made for Hungarian is that the definite article might additionally co-occur with an element quantifying a noun only if the noun is further restricted by a modifying element.This is shown in (30 a, b) for relative clauses and in (30 c, d) for possession.In (30 a, b) we can see that the phrase-initial definite article appears only if a relative clause is present between the article and the quantifier.(30 c, d) are the possessive counterparts.Az might appear only if a possessor phrase is present as well, which like in (30 a, b) for the relative clause, is again interposed between the definite article az 'the', and the quantifier minden 'every'.In a parallel fashion, except for the position of the modifier, in English a definite article must co-occur with a quantifier only if the noun is modified by a possessor phrase (ibidem: 19):

Old French, Old Sicilian and Old Italian
To the best of my knowledge, we have no texts attesting early uses of Apulian prepositionless genitives of the type of San Marco in Lamis, but we can look at other old Romance varieties employing such constructions found in works such as Jensen (2012Jensen ( [1990]]) and Delfitto and Paradisi (2009).In the latter work prepositionless genitives of Southern Italy are actually seen as a surviving instance of juxtaposed genitives of the Medieval Romance varieties they discuss.
Old French is a notable case.In Old French, in fact, possession could be expressed through the juxtaposition of two nouns, with genitives often being proper names.The genitive followed his head, as in (32) (data from Jensen 2012Jensen [1990]] Hence a first difference between the Apulian type of San Marco in Lamis and Old French is the possibility for proper names to appear as genitives in the relevant construction, contrary to what we have seen in (4).Jensen (2012) states in fact that the construction was mostly characterized by proper names, kinship terms, and high rank referents.It thus was a construction mostly dedicated to highly ranked referents, even though it was employed to express general ownership, too.
Proper nouns, which are articleless, were the ones mostly occurring in the construction, but articled ones were allowed too, as in (33) from Jensen (ibidem: 19): (33) dou pouoir l'Anemi gité from.thepower the enemy freed 'freed from the power of the enemy' Thus we also find instances nearly identical to those of the Apulian prepositionless genitives we discuss here, as in la corz lo roi, 'the King's court'.The construction was "used widely with nouns denotating human beings provided that the reference is to specific individuals and not to a class or category of people" (ibidem).So the nouns entering in the construction had to be [+specific] and [+human].The Apulian construction does not place any constraint on the [human] features of the two nouns even though it is dedicated to specific ones.Sometimes in Old French the order possessee-possessor would be reversed, yielding a genitival construction of the type selonc la Dieu benivolance, 'according to God's benevolence'; a la rei curt, 'to the King's court' (ibidem: 20), like the Old Italian la Dio mercé 'God's mercy'.
Prepositional genitives of the a type were instead usual with plural or generic possessors, se por le pechié as gens ne fust 'if it wasn't for people's sins'; ja mes n'entrera puis hui en chamber a dame n'a pucele 'never again from today will somebody enter a lady or a maiden's room' (ibidem).Prepositional genitives were thus dedicated to non-specific nouns, and to nouns headed by indefinite articles, and this extended to both the head and the argument noun (ibidem: 25): (34) a. la maison a une veve femme the house to a widow woman 'the house of a widow' b. une maison a un hermite trova a house to a hermit found 'he came upon the house of a hermit' In (34), we see how Old French follows the pattern of the Apulian and Romanian data in the preceding sections.Arteaga (1995) argues that the emergence of obligatory subject personal pronouns during the Middle French period and the disappearance of the juxtaposition genitive are linked to the loss of an Agr head, so that genitive case could not be assigned without an overt marker.This is suggested in Arteaga (1995: 87), whose syntactic representation of the construction is reported in (35):  As in general in the dialects of Southern Italy, we find that the Apulian variety of San Marco in Lamis is a pro-drop language.But prepositionless genitives still do not carry any genitival marker.
Other proposals suggesting null heads include Simonenko (2010: 9), where a treatment of genitival arguments as Kase projections occupying the complement position is envisaged, proposing a structure in which Kase is a silent head, occupied by the preposition in case of prepositional genitives.Given that modifiers must be human and specific, she considers the K head to be endowed with such features as well: isolation as further support for the theory that the language had an AGR head, discharging case to the complement noun to its right.A definiteness feature is considered necessary in order to assign case in Old French juxtaposed genitives by Delfitto and Paradisi (2009) as well.They propose an analysis à la Kayne in which the Agr head inherits its definiteness features from the possessor via spec-head agreement, as in Delfitto and Paradisi (2009: 60): In their theory the incorporation of Agr into D is triggered through a +human feature which syntactically activates AGR/K°.In doing this, the authors are trying to account for the fact that the construction restricts modifiers to human referents.
Old French prepositionless genitives have also been shown not to be iterable (like Old Italian ones).
The Apulian construction of San Marco in Lamis does allow iteration.Despite the fact that it obeys a several number of constraints on its realization like other types of prepositionless genitives do, the one on multiple occurrences does not seem to be one of them: (39) la kasa lu figgiə lu skarparə the house the son the cobbler 'the house of the cobbler's son' As we said earlier we have no data on this variety from around the period of Old French, on when it emerged in this precise variety, regarding whether it obeyed similar constraints to those of Old French, or whether they changed over time.The first written records appear towards the end of the 19 th century.On the other hand, Old Italian and Old Sicilian records containing a prepositionless genitive go back to at least the 14 th century.Possessors were in the form of proper names (unlike this Apulian variety, which does not allow them), and both the head and the argument had to be definite, as in the following examples quoted in Delfitto and Paradisi (2009: 63)  Finally, Delfitto and Paradisi see the construction as the one breaking the ground for Contemporary Italian N+N compounds.
The syntactic treatments of prepositionless genitives seen until now either include an empty Kase position for assigning genitive case, or incorporation of Agr into D, following the insight that definiteness is a common feature of this genitival construction.As definiteness does not seem to suffice in order to license it (remember the +human feature of Old French, for example), some scholars have tried to map further features on the K head.
Syntactic treatments of morphologically unmarked genitives generally try to answer the question about what, exactly, is the syntactic cue that allows speakers to interpret the argument as genitival even though it has no such a marking.As we said, when it comes to Old French some scholars considered the cue to be an Agr head whose disappearance is caused by the end of pro-drop subjects and of juxtaposed genitives (Arteaga 1995).Further research showed instead the lack of such a link (Delfitto and Paradisi 2009;Simonenko 2010), either because free and juxtaposed genitives coexisted, or because corpus studies showed that around the time juxtaposed genitives disappeared, 25% of the matrix clauses used in the dataset still had null subjects (Simonenko 2010: 5).Similarly, the Apulian variety under analysis here is a pro-drop language, has no morphological case, but allows morphologically unmarked genitives anyway.
Let us consider again that in Sammarchese the head of the construction might also be articleless as with vocatives, and heads as proper names (the latter pattern is also found in the Calabrian construction analyzed in Silvestri 2012).With this in mind and data such as (4), this means that it is possible to have a case of the type , but this pertains code mixing with Italian): the complex is interpreted as genitival only if the genitive DP is articled.We shall not draw a precise parallel with the cases we discussed in § 2 (remember genitive case in Afro-Asiatic, and Chinese, Balkan, and Indo-Iranian linkers), however given the role of the definite article which precedes the genitive in the interpretation of the latter as genitival, we conjecture that the pre-genitival article might be an obligatory element whose occurrence is necessary to allow the second DP to be interpreted as a genitive/possessor.In short, the inclusion relation is instantiated only when genitive nouns are headed by overt definite articles.

Conclusions
In the present work we have drawn a first sketch of prepositionless genitives in the Apulian variety of San Marco in Lamis.They have been compared with Modern Romance varieties (Romanian, Aromanian) and with Hebrew.Comparison with Old Romance varieties (Old French, Old Italian, and Old Sicilian) prepositionless genitives resulted in some similar patterns, but also in some discrepant ones, even though they all exhibit a constraint on the definiteness of the nouns entering the complex.In particular, for this Apulian variety we discussed a ban on heads followed by post-nominal adjectives, and on genitive nouns in the form of proper names (unlike Old French, Old Italian, and Old Sicilian, where genitives were mostly proper names).Post-nominal adjectives have been shown to be sitting only to the right of the genitival complex, never intervening between the head and the argument, and to be agreeing with the argument, rather than with the head.
The construction has shown to be iterable, and its arguments disallowed from appearing in isolation.Arguments may function as either subjects, or objects.Given its necessary presence, the article preceding the argument is understood as the element allowing the argument to be interpreted as the possessor in the inclusion relation.The following table recapitulates the patterns and the constraints which characterize the Apulian construction analyzed until now.
's photo' b. lu rətrattə də Lelina the portrait of Lelina 'Lelina's photo' son of the other king' c. lu figgiə dəll atu rre the son of.theother king 'the son of the other king' of the broken finger' ), acest obicei al vein-ului 'this habit of the neighbor' an al genitive, she envisages a structure in which D is occupied by a demonstrative or an indefinite article.The argument is thus generated as an adjunct to DP, comprising the linker al: every whim b. the every whim[of Margaret Tatcher] 5 [ DP ±raised] [ DP -raised] but not of the type *[ DP ±raised] [ DP +raised].*la madrə Fabio or *Anna Fabio are not possible sentences.Genitival constructions which in Italian are [ DP +raised] [ DP +raised] like casa Rossi are instead [ DP -raised][ DP -raised] in Sammarchese, la kasa li Sərrətellə, 'the house of the Serritellis' (speakers might employ the Italian [ DP +raised] [ DP +raised] : Like the Italian examples in ( b, c).So unlike in prepositionless genitives, both in al and non-al genitives case is morphologically expressed.Examples are from Dobrovie-Sorin et al. (2013: 314): : ): ): : ): from 14 th century Old Tuscan (40) and from 14 th century Old Sicilian (41):

Table 1 .
Lack of P pattern