In linguistics, a broken plural (or internal plural) is an irregular plural form of a noun or adjective found in the Semitic languages and other Afroasiatic languages such as Berber. Broken plurals are formed by changing the pattern of consonants and vowels inside the singular form. They contrast with sound plurals (or external plurals), which are formed by adding a suffix, but are also formally distinct from phenomena like the Germanic umlaut, a form of vowel mutation used in plural forms in Germanic languages.
There have been a variety of theoretical approaches to understanding these processes and varied attempts to produce systems or rules that can systematize these plural forms. However, the question of the origin of the broken plurals for the languages that exhibit them is not settled, though there are certain probabilities in distributions of specific plural forms in relation to specific singular patterns. As the conversions outgo by far the extent of mutations caused by the Germanic umlaut that is evidenced to be caused by inflectional suffixes, the sheer manifoldness of shapes corresponds to multiplex attempts at historical explanation ranging from proposals of transphonologizations and multiple accentual changes to switches between the categories of collectives, abstracta and plurals or noun class switches.
While the phenomenon is known from several Semitic languages, nowhere has it become as productive as in Arabic.
In Arabic, the regular way of making a plural for a masculine noun is adding the suffix -ūn[a] (for the nominative) or -īn[a] (for the accusative and genitive) at the end. For feminine nouns, the regular way is to add the suffix -āt. However, not all plurals follow these simple rules. One class of nouns in both spoken and written Arabic produce plurals by changing the pattern of vowels inside the word, sometimes also with the addition of a prefix or suffix. This system is not fully regular, and it is used mainly for masculine non-human nouns; human nouns are pluralized regularly or irregularly.
Broken plurals are known as jam‘ taksīr (جَمْعُ تَكْسِيرٍ, literally "plural of breaking") in Arabic grammar. These plurals constitute one of the most unusual aspects of the language, given the very strong and highly detailed grammar and derivation rules that govern the written language. Broken plurals can also be found in languages that have borrowed words from Arabic, for instance Persian, Pashto, Turkish, Kurdish, Azerbaijani, and Urdu. Sometimes in these languages the same noun has both a broken plural Arabic form and a local plural. E.g. in Pashto the word for "purpose" (مطلب) matlab can be pluralised in either its Arabic form مطالب matālib for more formal, High Pashto, or the plural مطلبونه matlabūna in everyday speech. (Cf. the treatment of Latin words in English; e.g. the plural of index is either indices or indexes, the latter being more informal.)
In Persian this kind of plural is known by its Arabic term jam'-e mokkassar (جَمِع مُکَسَر, literally "broken plural"). However the Persian Academy of Literature (Farhangestan) does not recommend the usage of such Arabic plural forms, but instead the native Persian plural suffix "-ha".
Full knowledge of these plurals can come only with extended exposure to the Arabic language, though a few rules can be noted. One study computed the probability that the pattern of vowels in the singular would predict the pattern in the broken plural (or vice versa) and found values ranging from 20% to 100% for different patterns.
A statistical analysis of a list of the 3000 most frequent Arabic words shows that 978 (59%) of the 1670 most frequent nominal forms take a sound plural, while the remaining 692 (41%) take a broken plural. Another estimate of all existing nominal forms gives over 90,000 forms with a sound plural and just 9540 with a broken one. This is due to the almost boundless number of participles and derived nominals in "-ī", most of which take a sound plural.
Semitic languages typically utilize triconsonantal roots, forming a "grid" into which vowels may be inserted without affecting the basic root.
Here are a few examples; note that the commonality is in the root consonants (capitalized), not the vowels.
- KiTāB كِتَاب "book" → KuTuB كُتُب "books"
- KāTiB كَاتِب "writer, scribe" → KuTTāB كُتَّاب “writers, scribes"
- maKTūB مَكْتُوْب "letter" → maKāTīB مَكَاتِيب "letters"
- maKTaB مَكْتَب "desk, office" → maKāTiB مَكَاتِب "offices"
- note: these four words all have a common root, K-T-B ك – ت – ب "to write"
In the non-semitic Persian language it is current to use:
- KiTāB كِتَاب "book" → KiTāBha "books"
- KāTiB كَاتِب "writer, scribe" → KāTiBha "writers, scribes"
Patterns in Arabic
|CiCāC||CuCuC||كِتَاب||kitāb (book)||كُتُب||kutub (books)|
|CaCīCah||سَفِينَة||safīnah (ship)||سُفُن||sufun (ships)||juzur (islands),
|CaCīC||سَبِيل||sabīl (path)||سُبُل||subul (paths)||turuq (paths)|
|CuCCah||CuCaC||غُرْفَة||ġurfah (room)||غُرَف||ġuraf (rooms)|
|CaCCah||شَقَّة||šaqqah (apartment)||شُقَق||šuqaq (apartments)|
|CiCCah||CiCaC||قِطّة||qiṭṭah (cat)||قِطَط||qiṭaṭ (cats)|
|CiCC||CiCaCah||هِرّ||hirr (cat)||هِرَرَة||hirarah (cats)||fiyalah (elephants)
|CvCC||CuCūC||قَلْب||qalb (heart)||قُلُوب||qulūb (hearts)||funūn (arts), buyūt (houses)
|عِلْم||ʿilm (science)||عُلُوم||ʿulūm (sciences)|
|جُحْر||juḥr (hole)||جُحُور||juḥūr (holes)|
|CvCC||CiCāC||كَلْب||kalb (dog)||كِلَاب||kilāb (dogs)|
|ظِلّ||ẓill (shadow)||ظِلَال||ẓilāl (shadows)|
|رُمْح||rumḥ (spear)||رِمَاح||rimāḥ (spears)|
|CaCaC||جَمَل||jamal (camel)||جِمَال||jimāl (camels)|
|CaCuC||رَجُل||rajul (man)||رِجَال||rijāl (men)|
|CvCC||ʾaCCāC||يَوْم||yawm (day)||أَيَّام||ʾayyām (days)||ʾarbāb (masters)
|حِلْم||ḥilm (prudence)||أَحْلَام||ʾaḥlām (meaning minds)|
|رُبْع||rubʿ (quarter)||أَرْبَاع||ʾarbāʿ (quarters)||ʾaʿmaq (deeps)|
|CaCaC||سَبَب||sabab (cause)||أَسْبَاب||ʾasbāb (causes)||ʾawlād (boys),
|CaCaCah||وَرَقَة||waraqah (paper)||أَوْرَاق||ʾawrāq (papers)||ʾašjār (trees)|
|CaCūC||ʾaCCiCah||عَمُود||ʿamūd (pole)||أَعْمِدَة||ʾaʿmidah (poles)||Ends with taʾ marbuta|
|CaCīC||ʾaCCiCāʾ||صَدِيق||ṣadīq (friend)||أَصْدِقَاء||ʾaṣdiqāʾ (friends)|
|CaCīC||CuCaCāʾ||سَعِيد||saʿīd (happy)||سُعَدَاء||suʿadāʾ (happy)||wuzarāʾ (ministers)||mostly for adjectives and occupational nouns|
|CāCiC||CuCCāC||كَاتِب||kātib (writer)||كُتَّاب||kuttāb (writers)||ṭullāb (students)
|Gemination of the second root; mostly for the active participle of Form I verbs|
|CāCiCah||CawāCiC||قَائِمَة||qāʾimah (list)||قَوَائِم||qawāʾim (lists)||bawārij (battleships)|
|CāCūC||CawāCīC||صَارُوخ||ṣārūḫ (rocket)||صَوَارِيخ||ṣawārīḫ (rockets)||ḥawāsīb (computers)|
|CiCāCah||CaCāʾiC||رِسَالَة||risāla (message)||رَسَائِل||rasāʾil (messages)||biṭāqah baṭāʾiq (cards)|
|CaCīCah||جَزِيرَة||jazīrah (island)||جَزَائِر||jazāʾir (islands)||haqāʾib (suitcases),
|CaCCaC||CaCāCiC||دَفْتَر||daftar (notebook)||دَفَاتِر||dafātir (notebooks)||applies to all four-literal nouns with short second vowel|
|CuCCuC||فُنْدُق||funduq (hotel)||فَنَادِق||fanādiq (hotels)|
|maCCaC||maCāCiC||مَلْبَس||malbas (apparel)||مَلَابِس||malābis (apparels)||makātib (offices)||Subcase of previous, with m as first literal|
|maCCiC||مَسْجِد||masjid (mosque)||مَسَاجِد||masājid (mosques)||manāzil (houses)|
|miCCaCah||مِنْطَقَة||minṭaqah (area)||مَنَاطِق||manāṭiq (areas)|
|CvCCv̄C||CaCāCīC||صَنْدُوق||ṣandūq (box)||صَنَادِيق||ṣanādīq (boxes)||applies to all four-literal nouns with long second vowel|
|miCCāC||maCāCīC||مِفْتَاح||miftāḥ (key)||مَفَاتِيح||mafātīḥ (keys)||Subcase of previous, with m as first literal|
|maCCūC||مَكْتُوب||maktūb (message)||مَكَاتِيب||makātīb (messages)|
In Hebrew, though all plurals must take either the sound masculine (-īm ־ים) or feminine (-ōt ־ות) plural suffixes, the historical stem alternations of the so-called segolate or consonant-cluster nouns between CVCC in the singular and CVCaC in the plural have often been compared to broken plural forms in other Semitic languages. Thus the form malkī מַלְכִּי "my king" in the singular is opposed to məlāxīm מְלָכִים "kings" in the plural.
In addition, there are many other cases where historical sound changes have resulted in stem allomorphy between singular and plural forms in Hebrew (or between absolute state and construct state, or between forms with pronominal suffixes and unsuffixed forms etc.), though such alternations do not operate according to general templates accommodating root consonants, and so are not usually considered to be true broken plurals by linguists.
Broken plurals were formerly used in some Ethiopic nouns. Examples include ˁanbässa "lion" with ˁanabəst "lions", kokäb "star" with kwakəbt "stars", ganen "demon" with aganənt "demons", and hagar "region" with ˀahgur "regions". Some of these broken plurals are still used in Amharic today, but they are generally seen as archaic.
- Ratcliffe, Robert R. (1998). The "Broken" Plural Problem in Arabic and Comparative Semitic. Current Issues in Linguistic Theory 168. Amsterdam/Philadelphia: John Benjamins. ISBN 9027236739.
- An overview of the theories is given by Ratcliffe, Robert R. (1998). The "Broken" Plural Problem in Arabic and Comparative Semitic. Current Issues in Linguistic Theory 168. Amsterdam/Philadelphia: John Benjamins. pp. 117 seqq. ISBN 9027236739.
- Ratcliffe, Robert R. (1998). The "Broken" Plural Problem in Arabic and Comparative Semitic. Current Issues in Linguistic Theory 168. Amsterdam/Philadelphia: John Benjamins. pp. 72–79. ISBN 9027236739.
- Boudelaa, Sami; Gaskell, M. Gareth (21 September 2010). "A re-examination of the default system for Arabic plurals". Language and Cognitive Processes. 17 (3): 321–343. doi:10.1080/01690960143000245.
- "Ge'ez (Axum)" by Gene Gragg in The Cambridge Encyclopedia of the World's Ancient Languages edited by Roger D. Woodard (2004) ISBN 0-521-56256-2, p. 440.
- “Hebrew” by P. Kyle McCarter Jr. in The Cambridge Encyclopedia of the World's Ancient Languages edited by Roger D. Woodard (2004) ISBN 0-521-56256-2, p. 342.
- Leslau, Wolf (1991). Comparative Dictionary of Geʿez (Classical Ethiopic). Wiesbaden: Harrassowitz, pp. 64, 280, 198, 216