This volume contains a selection of papers from the Second North American Conference in Iranian Linguistics (NACIL2) held at the University of Arizona in 2019. It starts with an Introduction that summarizes the main characteristics of Iranian languages, followed by a helpful review of Iranian linguistics to date. The Introduction also contains concise summaries of each of the eleven articles that make up the collection. For this reason, I will forgo doing the same in favour of highlighting, for each paper, one way in which it represents an advance in the thinking around the topic it addresses. Often the main argument or analysis in a paper rests on a key piece of data. In providing these examples in this review, I will retain the diverse transcription and glossing systems, as well as the translations, used by the authors. These include differences in the way vowels are represented ([u] vs. [oo] or [æ] vs. [a]) and the way inflectional elements like -rā (om for object marker vs acc for accusative case) and -i (ad for antidefinite vs. indf for indefinite) are glossed. The decision by the editors not to standardize the conventions used will not be confusing for those who work in the field and who are used to variation in transcription and glossing. Unless otherwise noted, the examples in this review are from Standard Persian.

Starting with advances that involve methodology, there are two papers that use experimental research to identify or verify empirical generalizations. In their paper on bare nominal objects in Chapter 1, Modarresi and Krifka (henceforth M&K) make an important contribution to the literature by showing that bare objects allow for anaphoric uptake. Bare objects in Persian are nominals that lack either overt indefinite marking or the differential object marker -rā. It has long been assumed that these bare objects are non-referential, meaning that they cannot serve as an antecedent for anaphoric reference later in the discourse. M&K show to the contrary that they can so serve based on five different experiments. For example, in one experiment, participants were asked to complete a sentence containing a bare noun, such as the one shown in (1). In 64% of the cases, they produced continuations that took up the antecedent such as referring back with a singular pronominal clitic as shown in (a) or a plural quantificational DP as in (b):

While M&K do not comment further on (1b), it is noteworthy that within the same sentence, the first instance of anaphoric uptake is plural while the second instance (bā-hash ‘with it’) is singular, indicating that there is more to explore here. The authors go on to provide a formal syntactic and semantic analysis of bare nouns as weak definites that is compelling.

The last chapter in this volume, by Storoshenko and Nakhaei, Chapter 11, similarly reports on cleverly designed experiments that seek to resolve a question about whether Persian has Negative Polarity Items (NPIs). Specifically, the authors explore whether elements involving hi:č ‘any’ are NPIs, i.e. items that must be licenced in the scope of negation, or whether they are subject to a weaker clausemate negation condition. The crucial case is shown in the following examples. In (2a) negation and hi:čkæs ‘anyone’ are in the same clause but hi:čkæs, as the subject, is presumably not in the scope of negation. In (2b) negation is in the matrix clause and hi:čkæs is in an embedded clause, where it is c-commanded by negation.

The results of their experiments confirm that (2a) is grammatical while (2b) is not. This supports the clausemate negation condition and entails that hi:č elements cannot be used to identify the position of negation in Persian, since a relationship with a c-commanding negative element is not required for their use. The chapter contains a clear and detailed description of the method used and a good discussion of how best to characterize hi:č elements, if not as NPIs. It ends with an intriguing observation that all languages that have elements that require clausemate negation, rather than c-commanding negation, appear to be verb-final.

There are two papers that use scope ambiguity to answer questions about syntactic structure: Chapter 2 on the scope of quantifiers by Shafiei and Chapter 3 on why-stripping by Rasekhi. Shafiei explores whether Persian is a scope-rigid language. She shows that while vP-external scrambling (e.g. a direct object scrambled to the left of the subject) does not induce ambiguity, vP-internal scrambling (e.g. a direct object scrambled to the left of a PP indirect object) does. That is, a sentence like (3a) must be interpreted with surface scope, while one like (3b) can receive an ambiguous interpretation, albeit a weak one as she acknowledges:

Shafiei’s analysis of these facts is constraint-based which allows her to capture the way in which inverse scope is only weakly available in some constructions.

Scope ambiguity is also used by Rasekhi in Chapter 3 as a way to distinguish why-stripping from sluicing. Rasekhi provides evidence that both of these constructions involve clausal ellipsis; however, she points out that they differ from one another in two ways. First, sluicing leaves behind only a wh-phrase as a remnant while why-stripping additionally leaves a non-wh-phrase behind. Second, sluicing can involve any wh-phrase while why-stripping, as the name suggests, only involves why.

Rasekhi proposes that these two constructions have different derivations. For sluicing, the wh-phrase moves to the Spec of FocP. For why-stripping, it is the remnant that moves to the Spec of FocP and why is base-generated higher in Spec of InterrogativeP. This difference in derivation is motivated by whether the constructions exhibit scope ambiguity. Rasekhi shows that with sluicing why can have lower scope, resulting in a pair-list interpretation, whereas with why-stripping, why can only have high scope with respect to the quantifier every:

Rasekhi presents a similar argument from the scope interaction of why with negation to support the idea that why in the two constructions is in a different structural position (92). While the judgements in both Shafiei’s and Rasekhi’s papers are subtle, they nevertheless require explanation and both papers provide plausible ones.

There are two papers in this collection on semantics of verbs. One involves making a distinction among verbs based on mood selection and the other distinguishes two types of anti-causatives. Starting with the former, Nematollahi in Chapter 7 is concerned with the properties of matrix predicates that determine whether their clausal complements are subjunctive or indicative. Offering a more nuanced approach than previous accounts that have focused only on tense or the realis/irrealis distinction (181–182), Nematollahi also considers predicate type (eventive vs. propositional) and then explores predicates that can select both indicative and subjunctive moods. For example, she shows that the following doxastic predicate can take either an indicative or subjunctive complement:

Nematollahi notes that in Romance languages, the use of the indicative has been claimed to come with a presupposition of the speaker’s commitment (194) but that in Persian the indicative signals the attitude holder’s perspective. This is shown in the following example where the speaker can use indicative mood in the complement but also express doubt about the truth of the proposition:

Nematollahi ultimately bases her analysis of mood selection on the work of Schlenker (Reference Schlenker, Geerts, van Ginneken and Jacobs2005).

Kasraie Kilakjani’s paper in Chapter 10 focuses on anticausatives, i.e. verbs that, like passives, lack an overt agent, but unlike passives do not meet the tests for an implicit agent. Adopting the view that external arguments are introduced by a voice head, and that certain anticausatives, particularly those that are morphologically unmarked, lack this head, he goes on to show that some anticausatives in Persian exhibit two properties that ‘can be taken as evidence for the existence of a voice layer’ (269). These two properties are topicalization with -rā, which he says is canonically a property of an active predicate, and compatibility with the adverb sahvan ‘unintentionally’:

While this point, that Persian anticausatives involve a voice head, is motivated, some of the other claims in the paper could be better supported. For example, Kasraie Kilakjani distinguishes two classes of anticausatives: those that can be topicalized by -rā and those that cannot (270–272) but does not explain which verbs fall into which class. Similarly, the idea of an agentivity continuum in Persian (278), while appealing, is not fully developed.

There are two papers that advance our thinking on verbal morphology and that also highlight the role that phonological considerations have in morphosyntactic analysis. In Chapter 9, Okubo and Nomoto (henceforth O&N) start with an interesting empirical fact that has not received much attention in the theoretical literature on Persian, namely that the distribution of the independent and dependent copular forms in the affirmative indicative present tense is not parallel.

O&N contrast the independent form hast with what they propose, following others, to be a null verbal copula. Both forms appear with person suffixes signalling agreement with the subject. Hast has been traditionally understood to be an existential verb while the null copula appears elsewhere. However, as O&N show there are three non-existential contexts in which the independent form must be used. Hast is obligatory (i) after so-called focus particles, such as the additive as shown in (8a), (ii) when the predicate has undergone ellipsis (8b), and (iii) when the whole predicational copular construction involves concessive intonation (8c):

O&N’s analysis is a phonological one, based on the idea that sentential stress is assigned to the leftmost element in vP (Kahnemuyipour Reference Kahnemuyipour2009). When operations apply that result in a null complement within vP, such as focus movement or ellipsis, the independent copula is required to host stress. Similar considerations predict the use of the independent copula where concessive intonation is required.

In Chapter 8 Suleymanov discusses stem opposition in the verbs of a group of genetically related Iranian varieties known as Tat. The binary stem opposition in verbs that Suleymanov considers is a characteristic feature of most, if not all, Iranian languages whereby some tense/aspect/mood categories are formed on a ‘present’ stem and others on a ‘past’ stem. Suleymanov, recognizing that the stems seldom carry semantic meaning, refers to them as Stem 1 and Stem 2 instead. Based on an extensive corpus of published material, he documents the ways in which many of the stems have undergone morphological levelling whereby one of the stems comes to resemble the other one. He points out the curious fact that this levelling is evident in most Tat dialects even though ‘they have long ceased to form a continuum and are no longer in active contact with one another’ (228).

Suleymanov briefly discusses the strategies that some varieties have developed for addressing the ambiguity that results from stem generalization, i.e. where Stem 1 and Stem 2 are identical. For instance, the following example from the Balaxanı sub-dialect of Abşeron Muslim Tat shows the introduction of stress as a morphological marker. Where Stem 1 and Stem 2 differ, stress is marked on the stem (9a and b) but where they are identical, stress appears on the stem in the subjunctive (9c) and on the agreement marker in the preterite (9d).

Suleymanov’s chapter is one of several that expands the domain of study beyond Modern Standard Persian to incorporate historical or geographical perspectives, the others being the three chapters on the Ezafe construction, Chapters 4, 5, and 6. Chapter 4 focuses on the genesis of Ezafe in Middle Persian, Chapter 5 describes the way in which Ezafe interacts with definiteness in Western Iranian languages, and Chapter 6 presents a comparison of Ezafe in a Northern Kurdish language with a marker that may be profitably viewed as a type of Ezafe in Turkish.

Both Samiian and Larson (henceforth S&L) in Chapter 4 and Karim in Chapter 5 attribute to Kent (Reference Kent1953) the idea that Ezafe derives from Old Persian (OP) haya, which served both as a relative pronoun and as a definite article. They also agree that the indefinite marker -i in Modern or New Persian (NP) comes from Old Persian aiwa. S&L (120:28) provide the following table, taken from Estaji (Reference Estaji, Bubenik, Hewson and Rose2009:11), which shows the intermediate stages of change for each marker through Middle Persian (MP) and Early New Persian (ENP):

It is noteworthy that these two markers, the Ezafe and the indefinite, cross paths, so to speak, in their diachronic development and it raises the question, not addressed by the authors, of how their distinct functions were maintained in Early New Persian.

S&L provide a rich set of examples to show that Middle Persian Ezafe is an independent morpheme, that it forms a constituent with its following phrase, and that it plausibly functions as a preposition. Evidence for it being a free morpheme is shown in (11a) where it cooccurs with the indefinite marker, something that is not possible in Modern Persian. Evidence for it forming a constituent with its following phrase is shown in (11b) where a modifying adjective phrase has been displaced to the end of a sentence. Evidence for it being prepositional comes from its ability to govern oblique case, to host oblique pronominal enclitics, and to alternate with az ‘of/from’ in various contexts (122–123).

One note about this chapter: it uses X-bar terminology that may be confusing to younger scholars such as referring to possessors as complements (107) and to the indefinite suffix -ēw as a specifier (108).

In Chapter 5, Karim describes the interaction between Ezafe and (in)definiteness in Western Iranian languages, focusing particularly on adjectival modification. One of his goals seems to be to trace the etymological source for Ezafe, indefinite, and definite articles from the diverse ways in which they are expressed across the ten languages he considers. This task is made more complicated by the differences in the ways in which modification and definiteness are often bundled together. He shows, for instance, that Kurmancî Kurdish has one Ezafe marker, -ê, on head nouns and their modifying adjectives (12a), and a different one, -î, on a head noun that bears the indefinite article as shown in (12b&c). In contrast, Soranî Kurdish has one form, -e, in definite contexts (13a), and a different one, -î, on nouns that are either overtly marked for indefiniteness (13c) or that bear neither definite nor indefinite marking (13b).

Kurmancî

Soranî

Karim acknowledges in his conclusion that the rich variation he discusses in his chapter is by no means exhaustive. A potentially fruitful avenue for further exploration would be the interaction between definiteness and number as it has been shown that the marker he calls the definite article -(h)e in colloquial New Persian, which tends to attach to unmodified nouns, is singular in construal (see, for example, von Heusinger & Sadeghpoor Reference von Heusinger, Sadeghpoor, Balogh, Latrouite and Van Valin‚2020:128 and references therein).

In Chapter 6, Gündoǧdu and Erbasi (henceforth G&E) compare Ezafe in Kurmanji Kurdish to a marker that appears in compounding and possession structures in Turkish. They take the bold step of suggesting that the Turkish marker too is a type of Ezafe. The Ezafe marker in Kurmanji Kurdish occurs on nouns followed by adjectival and nominal modifiers, relative clauses, prepositional phrases, and possessors. In Turkish there is a form with a more restricted distribution, appearing on the head noun in compounding and possessive constructions which, in contrast to Kurmanji Kurdish, is head final. G&E provide the following examples to highlight these similarities and differences. In (14a) and (15a) we see a head noun bearing an Ezafe marker modified by what they call a noun attribute, and in (14b) and (15b) we see a head noun bearing an Ezafe marker appearing with a possessor. However, while in Kurmanji Kurdish we see Ezafe on a head noun modified by an adjective (14c), in Turkish this is not possible as shown in (15c):

Kurmanji Kurdish

Turkish

G&E provide an analysis of these facts using the idea of a phase-domain. However, this analysis depends on several unconventional assumptions about nominal structure. These include proposing that nominal phrases involve two DPs, assuming that QP, the projection hosting quantifiers and numerals, is higher than DP, and asserting that functional PPs attach higher than PPs headed by nominal prepositions.

Overall, this volume is a very welcome addition to the English-language literature on Iranian linguistics. In introducing new methodologies, considering diachronic and phonological sources for contemporary syntactic phenomena, and in comparing understudied Iranian languages not only with each other but with neighbouring non-Iranian languages, the papers in this volume represent innovative advances in the field. This trend is sure to result in even more diverse and stimulating research that is sure to appear in the future.