Post by Category : Persian Grammar

2. Morphology

Morphology is the branch of linguistics that (as opposed to phonology) is concerned with the elements of the language that carry meaning.

Basic principles of morphology are explained in this chapter, primarily with regard to their relevance for Persian grammar.

Contents

a. Morphemes

The morpheme is a phoneme string which represents the smallest element of the language that carries meaning.

In the following statement, one counts 20 morphemes at a first glance:

همین نوکر هندیه که پول از جیبِ ارباب‌ش دزدید، بیرون‌ش کردند؟

Iraj Pezeshkzad (20th and 21st Century AD)

Morpho_01_ENDE

In linguistics, morphemes are classified according to the following feature characteristics:

As the above example shows these two classifications are not congruent, all feature combinations therefore appear: There are free and bound lexical morphemes, as well as free and bound grammatical morphemes.

b. Classification of Persian Bound Morphemes

The following classification of bound morphemes displays a central role within the Persian grammar:

Bound morphemes from the upper example can be classified as follows (VR = verbal root, Af = affix, Cl = clitic):

Morpho_04_EN

b.a Verbal Root

In Persian, verbal roots do not appear as free morphemes in this role. They are associated with processes or certain states as precursors of non-finite verb forms:

bon_EN


A complete list of Persian verbal roots is available in the chapter List of Participles in Persian within this website.

Grammatical categories of verbal roots (lexical aspect and transitivity) are discussed in chapter 13.c.a.

b.b Affix

Affixes adopt a special position under morphemes. An affix is a bound grammatical morpheme appended to a so-called (morphological) stem, to either change its grammatical categories (= inflectional affix), or to modify either its semantic meaning or grammatical role (= derivational affix).

Stems can be single lexical morphemes, or morpheme strings which contain at least one lexical morpheme:

  • Stem /hend/ + affix /-i/ → /hendi/ هندی
  • Stem /hendi/ + affix /-hɒ/ → /hendihɒ/ هندی‌ها
  • Affix /hæm-/ + stem /ɒhæng/ → /hæmɒhæng/ هم‌آهنگ
  • Affix /nɒ-/ + stem /hæmɒhæng/ → /nɒhæmɒhæng/ ناهم‌آهنگ

In this website, the structure composed of the stem and its affix is called affixal formation.

An affix and its stem are always located within the same word ↓.

Affixes are classified according to their location to the stem:

  • Prefixes are set in front of the stem, e.g. /be-/ in /bekɒr/ بکار and /hæm-/ in /hæmkɒr/ هم‌کار.

    Most Old and Middle Persian prefixes (like /æ-/ in /æmordɒd/ امرداد, /ɒ-/ in /ɒræmidɒn/ آرمیدن and /pæʒ-/ in /pæʒmordæn/ پژمردن) have lost their semantic role in New Persian, and are not considered as prefixes any more. It seems that only the following (non-productive) prefixes have made the transition to New Persian:

    1. /bol-/ (= “much”, cf. the Greek πολύ- = “poly-“) generates adjectives from nouns: /bolkɒmæ/ بلکامه, /bolhævæs/ بلهوس
    2. /pɒ-/, /pɒj-/, /pɒd-/ and /pɒr-/ (= “against”, cf. the Greek παρά- = “para-“) generates new nouns from nouns:

      /pɒzæhr/ پازهر, /pɒjzæhr/ پایزهر, /pɒdzæhr/ پادزهر (= “antitoxin”)

      /pɒsæng/ پاسنگ, /pɒjsæng/ پایسنگ, /pɒrsæng/ پارسنگ (= “balance weight on scales”, see also the article of Dr. Ali Ashraf Sadeghi in مجلهء زبانشناسی, 39rd issue, page 1)

    Nevertheless, some new prefixes have been built up in this language by means of grammaticalization ↓ of adverbs and prepositions (e.g. /næ-/ in /nægoftæ/ نگفته and /bi-/ in /bikɒr/ بی‌کار). These prefixes are discussed in the chapter 19..

  • Suffixes are appended to the stem, e.g. /-zɒr/ in /kɒrzɒr/ کارزار and /-gær/ in /kɒrgær/ کارگر. They represent the vast majority of affixes in Persian, and are covered in the chapters 20. and 13.d.
  • A circumfix consists of two segments which surround the stem to fulfill the role of one single affix, like /hæm- ~ -i/ in /hæmvelɒjati/ هم‌ولایتی and /hæmʃɒgerdi/ هم‌شاگردی or /nɒ- ~ -i/ in /nɒmɒdæri/ نامادری.

    It is striking that a circumfix is only assumed if the two segments fulfill one single role together. Therefore, the prefix /næ-/ and the suffix /-id/ in /nænevisid/ ننویسید are not recognized as a circumfix, because each of them determines a different grammatical category of the verb.

    Persian circumfixes are handled in the chapter 19..

  • Interfixes connect two stems. The only Persian interfix is /-ɒ-/, which is discussed in chapter 9.b.a: /sær-ɒ-pɒ/ سراپا, /jek-ɒ-jek/ یکایک.

    Interfixes are normally grammaticalized ↓ conjunction, and they shoud not be mistaken for epentheses (like /æ/ in /mehræbɒn/ مهربان or /i/ in /ɒgɒhɒnidæn/ آگاهانیدن). Epentheses are phonemes or phoneme strings inserted between morphemes for phonotactical reasons. They do not carry meaning (consequently no morphemes), and their omission does not change the meaning or the grammatical function of the statement: /mehræbɒn/ ⇆ /mehrbɒn/, /ɒgɒhɒnidæn/ ⇆ /ɒgɒhɒndæn/

    By contrast, interfixes certainly have a semantic role.

  • Infixes are inserted in the stem; they are not noted in Persian.
  • Transfixes are multisectional affixes which are interlocked with the stem. Transfixes are not noted in Persian, and typically occur in Semitic languages. However, the Persian language is confronted with them primarily at the formation of the Arabic plural:

    Arabic lexical morphemes consist typically of three (or occasionally four) consonants. Depending on number of consonants these lexical morphemes are called triliterals (e.g. /skn/ س‌ک‌ن and /ʤrb/ ج‌ر‌ب) and quadriliterals (e.g. /drhm/ د‌ر‌ه‌م and /flsf/ ف‌ل‌س‌ف), respectively. In the transfixing, segments of the affix are inserted between these consonants:

    • stem /skn/ + transfix /ma-c1-aa-c2-ij-c3/ → /masaakijn/ (in Persian /mæsɒkin/) مساکین
    • stem /ʤrb/ + transfix /ta-c1-aa-c2-ij-c3/ → /taʤaarib/ (in Persian /tæʤɒreb/) تجارب
    • stem /drhm/ + transfix /c1-a-c2-aa-c3-i-c4/ → /daraahim/ (in Persian /dærɒhem/) دراهم
    • stem /flsf/ + transfix /c1-a-c2-aa-c3-e-c4-a/ → /falaasefa/ (in Persian /fælɒsefæ/fælɒsefe) فلاسفه

    These transfixes can be noted in 3.1.g Annex: Arabic Pluralizing Transfixes.

All Persian affixes are listed in the thesaurus of this website.

For the sake of completeness it should be noted that many linguists describe all morphemes which are not affixes as bases, whether they are bound or free, lexical or grammatical. This classification of the morphemes has been renounced in this website.

b.c Clitic

The clitics represent a characteristic of the Proto-Indo-European language which has maintained itself predominantly in the Greek and Indo-Iranian languages. A clitic is an unstressed or weakly stressed bound morpheme which leans on a neighbouring expression.

Clitics are classified according to their position:

There are both lexical and grammatical clitics:

All Persian clitics are listed in the thesaurus of this website.


Some Persian clitics seem to be interstages of grammaticalization, which is a process of language change in historical linguistics at which lexical morphemes gradually become bound and grammatical, partially losing their morphological and phonological independence. This becomes clear in the following examples:

  1. The archaic suppletive verb /hænd/ as a free lexical morpheme, e.g. in:

    از مرد خرد بپرس!، از ایرا

    جز تو به جهان خردوران هند /hænd/

    Nasir Khusraw (11th Century AD)

  2. The suppletive verb /-ænd/ as a lexical clitic, e.g. in:

    اکثرِ روستاهامان در این حال‌ند /hɒl-ænd/.

    Jalal Al-e Ahmad (20th Century AD)

  3. The verbal inflectional affix /-ænd/, e.g. in:

    هر روز دسته‌جمعی زن و مرد و دخترها در موستان انگور می‌چیدند /mi-ʧidænd/ و خوشه‌هایِ درخشان را در لولا یا صندوق‌هایِ چوبی می‌گذاشتند /mi-gozɒʃtænd/. بعد آن لولاها را می‌بردند /mi-bordænd/ کنارِ رودخانه.

    Sadeq Hedajat (20th Century AD)

The pluralized forms of the partitive pronouns /xæjl-i/ خیلی and /bæʔz-i/ بعضی (= /xæjl-i-hɒ/ خیلی‌ها and /bæʔz-i-hɒ/ بعضی‌ها) can serve as further examples: the article /-i/ is understood as a derivational suffix by modern Persian speakers in these cases, therefore the inflectional suffix /-hɒ/ is attached to the article (instead of */xæjlhɒ-i/ خیل‌هایی and */bæʔzhɒ-i/ بعض‌هایی).

c. Morphs and Allomorphs

Similarly to the relation between phoneme, phones and allophones, the morpheme is understood in linguistics as an abstract class of one or more morphs, and single morphs of a morpheme are called allomorphs.

The simple form of allomorphy can be noted in (both lexical and grammatical) morphs which have a common etymology, e.g.:

Additionally etymological not-related morphs can be classed as allomorphs of a morpheme, specifically if they serve the same function. It is typically the case with inflectional affixes if several morphs cause the same modification of the grammatical category, e.g. the inflectional affixes /-ɒn/ and /-hɒ/ signalizing the plural in /deræxtɒn/ درختان and /deræxthɒ/ درخت‌ها.

d. Zero Allomorph and Zero Morpheme

Zero allomorphs are a specific form of the allomorphs of a morpheme. They are allomorphs which have no phone, nor a grammatical role. The sign for the zero allomorph is /Ø/.

In the following examples, the allomorphs [-æʃ], -ɒ and [-Ø] can be noted as past inflectional suffixes in third person singular:

گران‌مایه دستور گفتش /goftæʃ/ به شاه:

«نبایدت رفتن به آن رزم‌گاه»

Ferdowsi (10th and 11th Century AD)

گفتم: «گره نگشوده‌ام زآن طرّه تا من بوده‌ام»

گفتا /goftɒ/: «منَ‌ش فرموده‌ام تا با تو طرّاری کند»

Hafez (14th Century AD)

پرس‌پرسان می‌کشیدش تا به صدر

گفت /goftØ/: «گنجی یافتم آخر به صبر»

Rumi (13th Century AD)

A morpheme is called zero morpheme if it has a zero allomorph as the only allomorph. In Persian, for example, the zero morpheme of the imperative inflectional affix in second person singular can be noted:

بخور /boxorØ/ تا توانی به بازویِ خویش!

که سَعیَ‌ت بود در ترازویِ خویش

Saadi (12th and 13th Century AD)

e. Words and Lemmas

Words are defined according to phonological features: They are morphemes or morpheme strings which become separated from each other by limit signals, such as breaks.

In the example given above, one can recognize the following eleven words:

Morpho_05_ENDE

As this example highlights, words are not at all functional units: Clitics are functional units in the same word with the neighbouring units. However, besides the grammaticalization into clitics, further historical processes have led to the fusion of formerly independent functional units to form words, mainly because the original semantic distinction has been rendered extinct. For example:

  • The archaic phrase ↓ /ɒs-i ɒb/ آسی آب (= “water mill”, with the enclitical conjunction /-i/) has not developed into the phrase /ɒs-e ɒb/ آسِ آب in modern varieties of Persian: It is understood and pronounced as a single word (/ɒs-i-ɒb/ آسیاب). In this process, the meaning of the morpheme /ɒs/ آس (= “mill”) is not in the original conscious of most speakers, so new phrases like /ɒs-i-ɒb-e bɒdi/ آسیابِ بادی (= “windmill”), /ɒs-i-ɒb-e dæsti/ آسیابِ دستی (= “quern”) and even /ɒs-i-ɒb-e ɒbi/ آسیابِ آبی (“water mill”) are formed and used.
  • In the sentence سربه‌سرش گذاشته‌اند, the expression /sær-be-sær-æʃ/ سربه‌سرش is handled by speakers as a single word, although it consists of two sentence constituents (/sær/ سر as a direct object and /be sær-æʃ/ به سرش as a directive adverbial).

Lexicographical word forms, which are used as items in reference books, are called lemmas. Please note the lemmas to the words in the example given above:

Morpho_06_EN

As this example shows, infinitives are lemmas of verbs.

In addition, all inflectional affixes must be deleted from the word to achieve its lemma:

/ʃælvɒrhɒ/ شلوارها → /ʃælvɒr/ شلوار

/zibɒtærin/ زیباترین → /zibɒ/ زیبا

f. Phrases

In grammar, a phrase describes a morpheme string which forms a uniform syntactic role. Every phrase therefore consists of one or several direct constituents. Phrases are classified as follows:

g. Phrase Structure Models

The following models are used for the illustration of the morphological structure of statements:

  1. The hierarchical indention model has the advantage that the statement is represented up to the level of its single morphemes:

  2. The bracket model is more compact; however, it could become difficult to find the right bracket couples off-hand. This model does not usually represent the morphological analysis of affixal formations:

    • [IP [NP [NP [DetP [Dhæm-] [Din]] [NP [NP [Nnævkær] [Ahendi]] [D-e]]] [Conke] [IP [Npul] [PP [Pæz] [NP [Nʤib] [Con-e] [NP [Nærbɒb] [N-æʃ]]]] [Vdozdid]]] [Nbirun] [N-æʃ] [Vkærdænd]]

  3. The morphological analysis in the tree model is more understandable. Also this model usually does not analyze the affixal formations internally:

PhraseTree