Save

Prefixes and suffixes in Afroasiatic

No easy -fix

In: Brill's Journal of Afroasiatic Languages and Linguistics
Author:
Itamar Kastner University of Edinburgh UK Edinburgh, Scotland

Search for other papers by Itamar Kastner in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0002-1038-1703
Open Access

Abstract

How can we tell whether an agreement feature will end up as a prefix, a suffix, or a combination of exponents? Research on Afroasiatic languages has identified a number of asymmetries which can be found between prefixes and suffixes. This short review considers these asymmetries, points out three cross-Semitic generalizations, and outlines the possible sources for them. Four kinds of theoretical explanations are evaluated: Syntactic, Morphological, Morphotactic and Morphophonological.

1 Introduction

One of the main questions uniting the contributions in this special issue is whether a given agreement marker will appear as a prefix, a suffix, an infix, or will be distributed across different affixes. This question is common to much research on Afroasiatic languages, with the current consensus being that some asymmetries can be found between prefixes and suffixes in these languages. But what are the possible sources for these asymmetries and how should they be derived? In this brief review I identify three asymmetries (Section 2), outline how different kinds of theories attempt to derive them (Section 3), and conclude with broader discussion (Section 4).

2 Three generalizations

2.1 Between conjugations: Prefixal conjugation and suffixal conjugation

The most common way of looking at the difference between agreement markers is by contrasting a “prefixal conjugation”, where at least some of the agreement markers are prefixes, with a “suffixal conjugation”, in which they are all suffixes. The examples from Ṣanʕani Arabic in (1)–(2) show the second person marker as a prefix in one conjugation and as a suffix in the other, with the plural number marker a suffix in both (Watson, 1993; Hewett, 2022). This is a morphosyntactic distinction, relying on the tense and aspect features of the verb: the prefixal conjugation is the nonpast one and the suffixal conjugation the past one.

(1) Prefixal conjugation

ti-

gambir

(-u:)

2

sit.NPST

(M.PL)

‘You (all) sit’

(2) Suffixal conjugation

gambar

-t

(-u:)

sit.PST

2

(M.PL)

‘You (all) sat’

A look at full paradigms shows this pattern clearly, as with the following examples from contemporary Modern Hebrew, (3), here in the verbal template XaYaZ—where X, Y and Z are placeholders for root consonants (Kastner, 2019). The past tense forms of the verb katav ‘wrote’ comprise the suffixal conjugation, containing only suffixes. The future (nonpast) forms have only prefixes, or person prefixes coupled with number suffixes. More conservative variants distinguish gender in the suffixes as well.

(3)

Past, XaYaZ

Future, XaYaZ

SG

SG

PL

1

katáv

-ti

katáv

-nu

ji-

xtóv

ni-

xtóv

2M

katáv

-ta

katáv

-tem

ti-

xtóv

ti-

xtev-ú

2F

katáv

-t

katáv

-tem

ti-

xtev-í

ti-

xtev-ú

3M

katáv

katv

-ú

ji-

xtóv

ji-

xtev-ú

3F

katv

-á

katv

-ú

ti-

xtóv

ji-

xtev-ú

The first generalization, then, is that there is a morphosyntactic difference between prefixes and suffixes: the values of tense or aspect determine whether some agreement features are realized as prefixes or suffixes, in different “conjugations”.

2.2 Within conjugations: Prefixes differ prosodically from suffixes

In the suffixal conjugation all affixes are suffixal, but when we have two affixes in the prefixal conjugation, how can we tell which will be a prefix and which will be a suffix? One morphosyntactic generalization is that person markers are uniformly prefixes, whereas number/gender markers are either fused onto the prefix or spelled out separately as a suffix (Harbour, 2008); this can be seen throughout the right-hand side of (3).

There is another generalization that can be drawn, however, and that is a phonological one: the prefixes in the prefixal conjugation are all heavier (prosodically) than the suffixes. In the prefixal conjugation of contemporary Hebrew as in (3), the prefixes are all minimal feet of the shape CV, whereas suffixes are single vowels. Once we look at a number of languages, we can say impressionistically that agreement prefixes in Afroasiatic are very often heavier than agreement suffixes, especially when the two are instantiated in the same paradigm.

A stronger version of this claim holds that there is always a phonological difference to be found between a prefix and a suffix, when both spell out the same phi-features (but differ in tense/aspect). To see what this might look like, we can compare affixes between conjugations. The 1PL marker in (3) was ni- as a prefix and -nu as a suffix, both CV. But it’s not the case that the two affixes have the same phonological footprint; the distinction between prefix and suffix has other phonological correlates, in that the stems take different shapes. The stem in (4) is shorter, with a prefix, than that in (5), with a suffix, even though both affixes spell out the same phi-features.

(4) Prefix, monosyllabic stem

ni-xtov

1PL.FUT-write.SMPL.FUT

‘We will write.’

(5) Suffix, disyllabic stem

ka.tav-nu

write.SMPL.PST-1PL.PST

‘We wrote.’

2.3 Across conjugations: Template markers are prefixes or initial infixes

An additional point, which finds its way less often into this conversation, relates to templatic morphology. In languages that show overt affixation for verbal templates, this morphology is predominantly prefixal and infixal: templatic affixes sometimes appear as a prefix and sometimes as an infix towards the beginning of the word, while the stem vowels might also depend on the template. Zukoff (2022) gives examples like those in (6) from Modern Standard Arabic (McCarthy 1981, 384, and see also Al Kaabi and Ntelitheos 2019 on Emirati Arabic). In all cases the “reflexive” -t- is part of the verbal template—rather than, say, the root or agreement marking—and in all cases it appears early on in the word.

(6)

Template

Mnemonic

Example

Gloss

VIII

reflexive

(i)ktataba

‘wrote, registered’

V

reflexive of causative

(takattaba)

(constructed example)

VI

reflexive of applicative

takaataba

‘wrote to each other’

X

causative of reflexive

(i)staktaba

‘wrote, made to write’

2.4 Generalizations and questions

With these points in mind, we can put forward the following three generalizations as relevant to the distinction between prefixes and suffixes in Afroasiatic languages, albeit for now mostly based on data from Semitic:

(7) Generalizations about prefixes and suffixes:

a. Different syntactico-semantic profile: roughly nonpast vs past and person vs number.

b. Different phonological profile: roughly CV vs C / C vs V / V: vs V.

c. Templates are not suffixal but prefixal or an infix.

The research questions that reflect these generalizations can be formulated as follows:

Q1

How robust are each of the three generalizations in a language family?

Q2

How robust are each of the three generalizations in a given language?

Q3

What assumptions does a comprehensive formal analysis require in order to derive these generalizations?

We should also keep in mind two typological patterns. The first is that languages have a general suffixing preference (e.g. Himmelmann 2014; Martin and Culbertson 2020): there are more suffixes than prefixes within a language and across languages. The second is a general form of the pattern we have already seen, according to which person is usually expressed as a prefix and number as a suffix when these markers are both affixal: “person left, number right” (Harbour, 2008; Müller, 2020). A few of the accounts surveyed below will derive this latter generalization, but it’s by no means restricted to Afroasiatic.

The next section provides an overview of formal analyses, with a focus on the generalizations in (7) and on deriving them as in Q3.

3 Theoretical approaches to the asymmetry

This section goes through four kinds of approaches to the issue of affixation in Afroasiatic (and in general). Since they have different goals, different starting points and different theoretical assumptions, they cannot be expected to each derive all of the generalizations in (7). It is exactly in this sense that there is “no easy fix” for issues of affixation in Afroasiatic.

Since our goal is to find common threads in different works, I will on occasion simplify the exposition considerably compared to the original literature, in the hope of identifying the major lines of investigation. We will go through what I call the Syntactic approach, the Morphological one, the Morphotactic one and the Morphophonological one. Even though these are not necessarily the names which their proponents would have chosen, they will help illustrate the range of starting points in our current discussion. All approaches share the assumption that affixes spell out distinct functional heads in a syntactic structure in one way or another.

3.1 Syntactic

The first approach is the one we’ll call Syntactic. It assumes that what determines whether a given syntactic head is spelled out as a prefix or a suffix is its syntactic position with respect to the verb. If an affix starts off higher than the verb and no movement occurs, the affix will undergo some kind of morphological lowering to the verb and be spelled out as a prefix. But if the verb raises to the affix or higher, then this affix will be a suffix. The tree in (8) shows the process in broad strokes, assuming that Person, Number and Gender are base-generated higher than the verb, which itself undergoes head movement and adjoins to Gender. The resulting hypothetical linear order in this case would be Person-Number-V-Gender.

(8)

This conceptualization usually relies on the notion of head movement in the syntax and might be familiar from work on the Construct State (Ritter, 1988), the Agr literature in comparative syntax (Pollock, 1989), and other work on Semitic morphology (Shlonsky 1989, 1997; Benmamoun 2000; Martinović 2019; Faust 2021; see also the contributions by Cacchioli and Shlonsky, this issue).

Like any theory, this family of approaches relies on a number of ancillary assumptions. The hierarchical structure would ideally be motivated independently—is there other evidence for Person dominating Number as syntactic nodes?—giving rise to additional predictions (Harbour, 2007). Likewise, there should be independent reasons for why some heads raise and others lower, and how high.

How well does this kind of theory fare in deriving our generalizations? It can easily encode the syntactic differences between the prefixal and suffixal conjugations by manipulating the relevant features on V and T/Asp. The fact that prefixhood and suffixhood track the phonology is essentially an accident. And figuring out the templates will depend on making explicit how they emerge; if they are based anywhere below GenderP, as is most likely, then something else would need to be said in order to ensure that the verb does not undergo head-movement above the template, wrongly rendering it a suffix.

3.2 Morphological

The next influential approach can be called morphosyntactic or simply Morphological, based primarily on the work of Harbour (2008, 2016). In this theory, phi-features are hierarchically organized in what could be envisioned as either syntactic structure (i.e. as nodes in the syntactic tree, like in the Syntactic theory) or morphological structure (i.e. as some internal representation of an agreement node). A feature bundle contains person, number and potentially gender features; some are spelled out as prefixes and some as suffixes, and sometimes the entire feature bundle is spelled out as one affix (like throughout the suffixal conjugation). The innovative proposal here is that when the bundle is split up into prefix and suffix, this is done according to its internal structure. In additional, linear order tracks hierarchical order: structural precedence translates into linear precedence, roughly like in the Syntactic approach.

Say that the basic hierarchy looks like in (9), with person dominating number (this ordering was based primarily on implicational universals).

(9)

In some cases, the bundle is spelled out as a single affix, as with Hebrew ni-xtov ‘we will write’ (prefix: 1PL.Fut ni-). Since agreement is higher in the tree, this bundle will be spelled out as a prefix (using “*” to indicate linearization):

(10)

But what about cases in which there is both a prefix and a suffix, as in Hebrew ti-xtev-u ‘you all will write’ (prefix: 2.Fut ti-, suffix: PL -u)? Then the proposal is that one Vocabulary Item spells out person and one spells out number (in Distributed Morphology terminology; Halle and Marantz 1993). The system wants to maintain three sets of precedence relations: (a) person appears before the verbal stem, (b) person appears before number, and (c) the established adjacency of person and the stem should not be disrupted. We can satisfy all three constraints if person is spelled out as a prefix and number as a suffix:

(11)

The difference between the prefixal and suffixal conjugations can then be derived by anchoring the feature bundle to different heads (V, T/Asp), depending on their values. This approach improves on the Syntactic one by separating affixation from head movement. If agreement can be prefixal or suffixal even if the verb moves as high up as C (Harbour, 2007), then movement of the verb to T cannot be the sole factor determining linearization (Harbour, 2008, 2016).

One of the welcome outcomes of this theory is that it derives the crosslinguistic generalization mentioned earlier, according to which person is a prefix and number is a suffix when the two appear together (unless they are fused into one affix). Like the Syntactic approach, the Morphological one is meant to be general, applying to cases beyond Afroasiatic too.

Looking at our generalizations, the difference between the prefixal and suffixal conjugations is again captured by encoding at some point that the relevant value of T/Asp is linked to a different placement of affixes. And like in the Syntactic theory, the phonological difference between prefixes and suffixes remains a (historical) accident. Incorporating the generalization on templates would require some additional thought, since the template should slot between T/Asp and V, which is not usually the case; but since preserving an established linear adjacency is only a preference, some version of a violable constraint could be worked out.

3.3 Morphotactic

The third approach has been developed in recent work by Hewett (2022), who calls it “modular” or Fission-based, though we will use the term Morphotactic. The name is meant to convey the intuition that linearization results from purely morphology-internal operations, which do not rely on any other part of the grammar (save for the T/Asp-based distinction between prefixal and suffixal conjugations).

This theory has two main components, assuming that agreement features are hosted on T/Asp. First, feature nodes might be Fissioned off into two separate exponents. And second, this Fissioning happens before local dislocation (metathesis), which in turn applies before Vocabulary Insertion. So while the Syntactic approach can be lexicalist, and while the Morphological approach assumes that the grammar first spells out the feature bundle or different parts of it and then arranges them, the Morphotactic approach arranges the abstract morphemes around the verbal stem first, and only then calls Vocabulary Insertion. As we’ve done with the previous approaches, we’ll outline how this theory works, discuss some of its predictions, and then evaluate it with respect to the generalizations in (7).

If agreement features originate on T, and a head-final clause arises due to successive head-movement, then we will always end up with the order V-Asp-T. This is in a sense the default: a suffixing paradigm. To see how prefixation occurs, let’s return to Ṣanʕani Arabic ni-gambir ‘we will sit’, with prefixed 1PL.FUT. First, assume a rule that metathesizes T and Asp in the language. This rule has the effect of placing T before the verbal complex, creating a prefix (Hewett, 2022, 21):

(12)

How do forms with both a person prefix and a number suffix arise, then, like in nonpast ti-gambir-ayn ‘you two will sit’? A rule of Fission applies first, “splitting up” all relevant feature bundles, in this case 2nd and 3rd person, into effectively one person morpheme and one number morpheme (Hewett, 2022, 19):

(13)

Applied to our example, the single T head undergoes Fission as follows (Hewett, 2022, 23):

(14)

Metathesis then applies as it did earlier, swapping the order of Asp (the verbal stem) and the closest T, here number. As a result, person ends up to the left and number to the right:

(15)

The viability of this approach is boosted by a prediction it makes for possible cases of contextual allomorphy. As Hewett (2022) explains, if morphemes (terminals) are first rearranged and only then spelled out, then the conditioning environments for allomorphy will be local in that prefixes would be conditioned by word-initial material and suffixes by word-final material. His survey of Semitic languages shows that this is the case. But if Vocabulary Insertion were to happen first, as in the Morphological account, we would wrongly predict that e.g. word-final material could condition the form of prefixes.

Reviewing now the assumptions of this account, the Fission rule in (13), while highly specific, does predict a pattern of “metafission” across different paradigms, thereby validating this specific mechanism. The metathesis rule needs to apply in order to encode the difference between the prefixal and suffixal conjugation.

Considering the generalizations to be accounted for, this theory easily derives the difference between past and nonpast conjugations by manipulating the value of T/Asp in its rule of metathesis. Like in the first two theories, the phonological profile of affixes is not addressed. No explicit mention is made of templates, so we can assume that they are generated either on v or on Voice. On some formal accounts, and certainly in many descriptions of templatic morphology, Semitic templates involve applicative and causative elements (Wallace, 2013; Al Kaabi and Ntelitheos, 2019; Kastner, 2020; Mashaqba et al., 2021; Zukoff, 2022). In most formal approaches, this kind of Appl or Caus should be below T/Asp. Yet these elements would all wrongly be predicted to be suffixal, if the theory subscribes to head movement, deriving a suffixing structure: V-Voice-template–Asp-agreement.

There are two possible solutions to this problem. The first is to separate agreement morphology from template-specific morphology: perhaps the former is linearized in the Morphotactic way and the latter another way (like in Section 3.4). This view separates the mechanisms deriving non-concatenative morphology (template-specific) from concatenative but perhaps discontinuous morphology (agreement).

The second is to assume additional rules of methathesis, such that all affixation is Morphotactic: if templates are generated on Voice, then v and Voice would need to swap places in order to give the order (prefixes-)V-Voice-Asp(-suffixes). Considering potential heads like Appl, a series of metathesis rules would need to be developed in order to make sure that the exponents of templates end up prefixal, regardless of whether they are below Voice, on it or above it. While such an undertaking is not out of the question, it remains to be further motivated. In the meantime, it does turn our attention towards the way that the templatic system is shaped—the starting point for the last set of theories reviewed here.

3.4 Morphophonological

Whereas in the first three kinds of theories we’ve seen, affixes are placed by the syntax, the morphosyntax or the morphology, the family of accounts we call Morphophonological differ in that placement of affixes is a matter dictated to a significant degree by the phonology. These theories span a range of frameworks and implementations, but all share the following assumptions (Tucker, 2010; Wallace, 2013; Bendjaballah, 2014; Kastner, 2019, 2020; Zukoff, 2022): the order of affixes mirrors their position in the syntax, there is no morphological hierarchy or markedness of features, and—most importantly in the current context—alignment of an affix with the stem is phonologically optimizing. It’s important to stress that the phonology does not do all the work of placing affixes, otherwise we would never be able to derive the difference between prefixal and suffixal conjugations. Instead, once an exponent is selected, the phonology is the one to dictate whether it is a prefix or a suffix.

What counts as “phonologically optimizing” differs by implementation. For Tucker (2010), the root needs to be aligned with the right edge of the prosodic word. For Wallace (2013), morphemes are aligned to edges of words. And for Kastner (2019), the grammar attempts to avoid complex onsets. Let’s take the implementation in Kastner (2019, 2020) as an example of these theories.

The syntax is set up uncontroversially for a generative approach within Distributed Morphology, with Voice as the locus of templatic morphology, including templatic affixes and stem vowels. Agreement is generated on T/Asp. Spell-out is cyclic, applying first to the VoiceP domain and then the TP domain. What this means is that stem vowels and templatic morphology are determined first, followed by agreement morphology. Vocabulary Insertion rules list the exponents of each morpheme (Vocabulary Items). A form without overt affixes, like Hebrew taraf ‘he devoured’, requires the VI s in (16); their optimal concatenation is that in (17a), and not any of (17b–f); I omit the actual constraints used to derive this result (Kastner, 2020, 46).

(16) Cycle 1, VI:

a. trf

b. v ↔ (covert)

c. Voice ↔ a,a / T[Past] ___

(17) Cycle 1, Phonology: trf + a,a =

a. taraf

b. *aa-trf

c. *trf-aa

d. *atraf

e. *tarfa

f. …

The feminine form, tarf-a ‘she devoured’, has an overt suffix. The relevant VI rule is given in (18) and the next steps in the cyclic spell-out in (19). A separate phonological process in the language syncopates the second stem vowel.

(18) Cycle 2, VI:

3SG.Fa / T[Past] ___

(19) Cycle 2, Phonology: a + taraf =

/a-taraf/ ⇒ /tar⟨a⟩f-a/ ⇒ tarfa (*ataraf, *atarf)

What happened here is that the affix, which happens to be a single vowel, was placed as a suffix by phonological considerations (whose details, again, we skip).

In other cases, the affix might have a different phonological profile, leading it to be placed as a prefix. For ji-gnov ‘he will steal’, the relevant VI s are in (20) and the phonological derivation proceeds as in (21).

(20) a. gnv

b. Voice ↔ o / T[Fut] ___

c. T[3SG.Fut] ↔ ji

(21) a. Cycle 1: o-gnv ⇒ /o-gnv/ ⇒ gnov (*gonv)

b. Cycle 2: ji-gnov ⇒ /ji-gnov/ ⇒ jignov (*gnovji)

Nothing special is going on when two affixes appear together, one as a prefix and one as a suffix: each is placed where it needs to be according to its phonological profile, which happens to work out just fine (as long as the two affixes are already stipulated in the VI s). In fact, this already happens even when there is only one overt agreement affix, since the templatic morphology—for example the stem vowel or vowels—is first placed as a prefix, infix or suffix. So in this kind of theory, there is no difference in the mechanism placing stem vowels, templatic affixes, person affixes or number affixes: the phonology finds the optimal arrangement for each one, going cyclically up the tree.

With the basics explained, a few points remain to be addressed. First, if the only difference between prefixes and suffixes is phonological, we might predict that a certain phonological profile (say CV) can only be one or the other. We saw earlier on, in (4)–(5) and repeated here, that this isn’t quite right: Hebrew has prefixal ni- in one conjugation and suffixal -nu in the other. But we also saw that the stems differ in the two conjugations, in that the former is monosyllabic and the latter disyllabic. So if the algorithm is phonological, then it might be the case that CV is an optimal prefix in one context but an optimal suffix in another.

(22) Prefix, monosyllabic stem

ni-xtov

1PL.FUT-write.SMPL.FUT

‘We will write.’

(23) Suffix, disyllabic stem

ka.tav-nu

write.SMPL.PST-1PL.PST

‘We wrote.’

The Morphophonological approaches also allow us to combine agreement morphology with patterns of stress and syncope, explaining the correlation between non-concatenative morphology and prosodic restrictions (McCarthy, 1979, 1981; Ussishkin, 2000, 2005; Kastner and Tucker, submitted). They further make a number of predictions regarding allomorphy which we have not explored here. However, since they order VI before placement, they cannot derive the generalization which set the Morphotactic account apart empirically.

Summing up from the vantage point of our generalizations, now the difference between prefixal and suffixal conjugations is not encoded in one rule. Rather, stems of the former have one phonological profile, and stems of the latter another phonological profile. The phonological difference between prefixes and suffixes is now the driving force of the analysis, arising as a result of the phonological calculation. And since templatic morphology is part of the same system, the placement of all exponents relating to the verbal template is correctly captured by the same system. We sum up the comparison next, and then return to some issues left implicit thus far.

3.5 Summary

The following generalizations were posed in (7) as challenges for any formal theory of affixation in Afroasiatic:

(24) Generalizations about prefixes and suffixes in Afroasiatic:

a. Different syntactico-semantic profile: roughly nonpast vs past and person vs number.

b. Different phonological profile: roughly CV vs C / C vs V / V: vs V.

c. Templates are not suffixal but prefixal or an infix.

Current theories differ primarily in whether they start from (24a) or (24b), as surveyed above. Either decision requires a series of assumptions which constrain the theory and make additional predictions, but while each has its strong suits, they have not yet converged.

4 Discussion

To round off the discussion of different approaches to affixation, we revisit some issues which have remained implicit so far, or which not all theories acknowledge explicitly.

The first is the nature of the asymmetries, and whether they should be treated as synchronic or diachronic. For example, is the difference between prefixal and suffixal conjugations something that should be encoded in the system, or is it a historical artifact which a formal system does not need to derive (cf. Embick 2010 for a related discussion of allomorphy as non-optimizing)? And what about the phonological difference between prefixes and suffixes, is it synchronic or diachronic? If we answer “yes” to one, we need to have good reason to answer “no” to the other, be that reason conceptual or empirical. At some point, at least one of these two forces shaped the Proto-Afroasiatic agreement system which its descendant languages inherited. A way of problematizing this question is by asking whether we want to have theories which could just as easily allow suffixes that are phonologically heavier than prefixes, or alternatively a prefixal conjugation for past and a suffixal conjugation for nonpast.

Another question is how much these theories should generalize beyond Afroasiatic. All of the theories surveyed above have found support in other language families. At what point do we think that the patterns under discussion are particular to Afroasiatic (or even Semitic), and how much of them do we think are more general? For example, given emerging generalizations about the morphophonology of infixation in general (Kalin, 2022), do we want to treat agreement prefixes in Afroasiatic, template prefixes/infixes in Semitic and other kinds of infixes all in the same way?

The typological question becomes clearer when seen through the lens of the generalizations mentioned earlier on, namely the suffixing preference and “person-left, number-right”. The importance of the latter has been part of our discussion throughout, but the suffixing preference also raises a number of important issues about processing, learnability and acquisition. None of the theories we mentioned here has made explicit claims on these topics which would set it apart from its peers. Perhaps that will be one way of converging on a comprehensive theory.

Acknowledgments

Thanks to Matthew Hewett, Ruth Kramer and an anonymous reviewer for comments, and to Mohamed Lahrouchi, Andrew Nevins, Ur Shlonsky and the participants of the March 2022 workshop on affixation in Afroasiatic, where a previous version of this work was presented. Thanks to Matthew Hewett also for providing the figures.

References

  • Al Kaabi, Meera, and Dimitrios Ntelitheos. 2019. Rethinking templates: A syntactic analysis of verbal morphology in Emirati Arabic. Glossa: A Journal of General Linguistics 4.

    • Search Google Scholar
    • Export Citation
  • Bendjaballah, Sabrina. 2014. Remarks on nonconcatenative affixation. Brill’s Annual of Afroasiatic Languages and Linguistics 6:4573.

    • Search Google Scholar
    • Export Citation
  • Benmamoun, Elabbas. 2000. The feature structure of functional categories: A comparative study of Arabic dialects. Oxford: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Embick, David. 2010. Localism versus globalism in morphology and phonology. Number 60 in Linguistic Inquiry Monographs. Cambridge, MA: MIT Press.

    • Search Google Scholar
    • Export Citation
  • Faust, Noam. 2021. Modern hebrew prefixoids: description and morpho-syntactic analysis. Morphology 32:2544.

  • Halle, Morris, and Alec Marantz. 1993. Distributed Morphology and the pieces of inflection. In The view from building 20: Essays in linguistics in honor of Sylvain Bromberger, ed. Ken Hale and Samuel Jay Keyser, 111176. Cambridge, MA: MIT Press.

    • Search Google Scholar
    • Export Citation
  • Harbour, Daniel. 2007. Against PersonP. Syntax 10:223242.

  • Harbour, Daniel. 2008. Discontinuous agreement and the Syntax–Morphology interface. In Phi theory: Phi-Features across modules and interfaces, ed. Daniel Harbour, David Adger, and Susana Béjar, 185220. Oxford: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Harbour, Daniel. 2016. Impossible persons. Cambridge, MA: MIT Press.

  • Hewett, Matthew. 2022. Allomorphy in semitic discontinuous agreement: Evidence for a modular approach to postsyntax. Natural Language & Linguistic Theory URL http://dx.doi.org/10.1007/s11049-022-09559-w.

    • Search Google Scholar
    • Export Citation
  • Himmelmann, Nikolaus P. 2014. Asymmetries in the prosodic phrasing of function words: Another look at the suffixing preference. Language 90:927960.

    • Search Google Scholar
    • Export Citation
  • Kalin, Laura. 2022. Infixes really are (underlyingly) prefixes/suffixes: Evidence from allomorphy on the fine timing of infixation. Language 98:641682.

    • Search Google Scholar
    • Export Citation
  • Kastner, Itamar. 2019. Templatic morphology as an emergent property: Roots and functional heads in Hebrew. Natural Language & Linguistic Theory 37:571619.

    • Search Google Scholar
    • Export Citation
  • Kastner, Itamar. 2020. Voice at the interfaces: The syntax, semantics and morphology of the Hebrew verb. Number 8 in Open Generative Syntax. Berlin: Language Science Press.

    • Search Google Scholar
    • Export Citation
  • Kastner, Itamar, and Matthew A. Tucker. submitted. Non-concatenative morphology. In The Cambridge handbook of Distributed Morphology, ed. Artemis Alexiadou, Ruth Kramer, Alec Marantz, and Isabel Oltra Massuet. Cambridge: Cambridge University Press.

    • Search Google Scholar
    • Export Citation
  • Martin, Alexander, and Jennifer Culbertson. 2020. Revisiting the suffixing preference: Native-language affixation patterns influence perception of sequences. Psychological Science 31.

    • Search Google Scholar
    • Export Citation
  • Martinović, Martina. 2019. Interleaving syntax and postsyntax: Spellout before syntactic movement. Syntax 22:378418.

  • Mashaqba, Bassil, Anas Huneety, Wael Zuraiq, Moh’D Al-Omari, and Sabri Al-Shboul. 2021. Labile anticausatives in Jordanian Arabic. Lingua Posnaniensis 62:1945.

    • Search Google Scholar
    • Export Citation
  • McCarthy, John J. 1979. Formal problems in Semitic phonology and morphology. Doctoral Dissertation, Massachusetts Institute of Technology.

    • Search Google Scholar
    • Export Citation
  • McCarthy, John J. 1981. A prosodic theory of nonconcatenative morphology. Linguistic Inquiry 12:373418.

  • Müller, Gereon. 2020. Inflectional morphology in harmonic serialism. Advances in Optimality Theory. Equinox.

  • Pollock, Jean-Yves. 1989. Verb movement, UG, and the structure of IP. Linguistic Inquiry 20:365424.

  • Ritter, Elizabeth. 1988. A head-movement approach to construct-state noun phrases. Linguistics 26:909929.

  • Shlonsky, Ur. 1989. The hierarchical representation of subject verb agreement. Ms., University of Haifa.

  • Shlonsky, Ur. 1997. Clause structure and word order in Hebrew and Arabic: An essay in comparative Semitic syntax. Oxford: Oxford University Press.

    • Search Google Scholar
    • Export Citation
  • Tucker, Matthew. 2010. Roots and prosody: the Iraqi Arabic derivational verb. Recherches Linguistique de Vincennes: Racine et Radical 39:3168.

    • Search Google Scholar
    • Export Citation
  • Ussishkin, Adam. 2000. The emergence of fixed prosody. Doctoral Dissertation, University of California, Santa Cruz, Santa Cruz, CA.

  • Ussishkin, Adam. 2005. A fixed prosodic theory of nonconcatenative templatic morphology. Natural Language & Linguistic Theory 23:169218.

    • Search Google Scholar
    • Export Citation
  • Wallace, Katherine. 2013. A concatenative approach to Semitic templatic morphology. Ms., NYU. lingbuzz/002299.

  • Watson, Janet C.E. 1993. A syntax of Ṣanᶜānī Arabic. Wiesbaden: Harrassowitz.

  • Zukoff, Sam. 2022. The mirror alignment principle: Morpheme ordering at the morphosyntax-phonology interface. Natural Language & Linguistic Theory.

    • Search Google Scholar
    • Export Citation

Content Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 1105 794 33
PDF Views & Downloads 1047 698 15