The present paper looks at issues concerning the noun phrase in Bashkardi. “Bashkardi” is taken here in the narrower sense, referring to the varieties spoken in the region of Bashakerd, inland from Bandar Abbas. Apart from two survey articles, and words and sentences quoted by GERSHEVITCH in his articles, these varieties are essentially undescribed. This particularly applies to nominal inflection, for which SKJÆRVØ (1989a: 848) merely states a plural marker on and the possible existence of an “indefinite article -o”. The data for this paper come from recordings made in Bashakerd by Ilya Gershevitch in 1956. This material is particularly interesting for the fact that it dates from a period before paved roads and electricity became available in the rural parts of Iran, so the influence of Persian was less strong than it is today, thus offering precious data for a diachronic study. In describing the Bashkardi noun phrase, I will specifically look out for details that might shed some light on the history of the nominal system. I will argue that the data show a system in a somewhat fluid state, mirroring a transition between the loss of inherited case markers and the grammaticalisation of new ones. 1. Number and case Nouns and pronouns are largely unmarked for case, and to some extent even for number. The generic noun may be used in plural function when indefinite: in (1), the statement translatable as ‘I have a sister’ also permits the reading ‘I have sisters’. The distinction of direct and oblique case has been lost, so the suffix -ān (originally OBL.PL) with variants (-ōn etc.) is the general PL marker (2) by a process parallel to the one that occurred in Persian (but differently from Balochi etc.). As in Balochi, mass nouns are conceived as plural (3). The “indefinite” -o(y) mentioned by SKJÆRVØ does occur in (some of) South Bashkardi, but is not very frequent; it is translatable in its probable etymological sense ‘one’ (2). Much more common is the clitic -ē/ -ī known from Persian and other Ir. languages (4). To a certain extent, adpositions and particles indicate syntactic relations, but they do not seem grammaticalised yet, and the unmarked noun is widely used in all syntactic functions. 2. Object marking A suffix -ā/a is occasionally found, which recalls the Bal. OBL.SG marker -ā. The only clear instance of -ā marking an object that I found so far is (5), where the position after the pronominal clitic is noteworthy. In all other instances, -ā marks direction or other adverbial meanings (7a). The combination of -ā with -ān does not occur in the data. This situation fits the opinion voiced by various authors for Balochi that the case marker -ā is a comparatively recent addition to the nominal system. Prepositions occur quite commonly in the PRESENT and PAST domains to mark direct and indirect objects. It seems that definiteness is a condition for such marking. Most instances are pronouns or otherwise refer to animates, but inanimates are also found, as in (3), where the water is definite as it was mentioned before. The same prepositions are also used in local functions (then mostly referring to inanimates). There is thus no dedicated object marker in Bashkardi, as in Middle Persian and Parthian. 3. Possessive constructions As in many other Ir. languages, an existential sentence is used to express ‘have’ (1). For genetive relations, both head-final and head-initial patterns are found. Head-initial ones occur with and without ezāfe. The latter pattern is rather regular with pronominal possessors (6), but also occurs elsewhere (7b). The unmarked noun occurring in genitive function corresponds to Middle Persian and Parthian patterns. Particularly interesting are instances with a suffix -ī on the possessor (7a), (4). Since this -ī is clearly unstressed, it cannot be the suffix found in Persian adjectives of appurtenance (e.g. Šīrāzī ‘from Shiraz’). It is unclear at present whether this -ī is a case suffix (possibly a genitive ending deriving from Old Ir. -ahya as does the Balochi genitive -ay / -ē ?), or a peculiar use of the pronominal clitic of the 3SG (in this case, one might rather expect its being suffixed to the possessum in the way of Turkic ‘a lion it’s kid’ or German dialectal meiner Schwester ihr Haus ‘to my sister her house’ = ‘my sister’s house’). Examples (1) K:xūx=etheš?‘Do you have sisters?’ sister=PC2SGexists A:xūx=omheš.(...) ‘I have sister(s).’ sister=PC1SGexists K:čanxūx=etheš?‘How many sisters do you have?’ how manysister=PC2SGexists A:dōxōx‘Two sisters.’ twosister(T I.3, 21ff., Jaġdān, SB) (2) hå,verx=oy=omdīx,båsīd-an=omdīx yesleopard=IND=PC1SGsee.PFwolf-PL=PC1SGsee.PF ‘Yes, I saw one leopard and [several] wolves.’ (T I.1, 26, Jaġdān, SB) (3) moneiyåw-ōna-xwar-om.‘I drink the water.’ Itowater-PLIPF-eat.PR-1SG (UC 3 T I.2, 18, Rameshk?, NB) (4) ravånbǖ,röuǰåh-ēdīst=īyačukšīr=īe. goingbecome.PTgo.PTplace=INDsee.PT=PC3SGonekidlion=ICOP.PR3SG ‘He set off [and] at a certain place, he saw: there is a lion kid.’ (T II.2, 34, Angohrān, NB) (5) xo,hålåpåbatåbe-rr-īn wellnowfootbecome.PR.IPR2SGso thatSBJ-go.PR-1PL tådehngōn=et-aseråkbe-dah-om. so thatchief=PC2SG-AshowingSBJ-do.PR-1SG ‘Now get up; let’s go [and] I will show you your chief.’ (T II.2, 126, Angohrān, NB) (6) esmxwad=omna-go‘I didn’t say my name [yet].’ nameself=PC1SGnot-say.PT(T IV.1, 64, SB) (7) a)tulag-a=īsīr-å‘[I will go] to the jackals’ wedding’ jackal-PL=Iwedding-A(C 4 T III.2, 42=48=54=58, NB) b)sīrtolaw-onar‘It was the jackals’ wedding.’ weddingjackal-PLCOP.PT.3SG(C 4 T III.2, 49, NB) References BARBERA, Gerardo 2005: Lingua e cultura a Minab (Iran sud-orientale). Profilo grammaticale, testi e vocabolario. Naples (unpublished PhD thesis) SKJÆRVØ, Prods O. 1989a: "Baškardi." In: Encyclopædia Iranica III, pp. 846-850 _____ 1989b: "Languages of Southeast Iran." In: Rüdiger SCHMITT (ed.): Compendium Linguarum Iranicarum. Wiesbaden: Reichert, pp. 363-369