{{Short description|Aspect of Sanskrit grammar}} Sanskrit inherits from its parent, the Proto-Indo-European language, the capability of forming compound nouns, also widely seen in kindred languages, especially German, Greek, and English.{{cn|date=August 2022}}

However, Sanskrit, especially in the later stages of the language, significantly expands on this both in terms of the number of elements making up a single compound and the volume of compound usage in the literature, a development which is unique within Indo-European to Sanskrit and closely related languages.{{cn|date=August 2022}}

Further, this development in the later language{{efn|"In Vedic, noun compounds are hardly more frequent than in Homeric Greek, but their frequency increases throughout the history of the language.. in the later language the occurrence in a single short sentence of several compounds of four or five members is perfectly normal, and in certain styles compounds of twenty or more members are not thought excessive" - Coulson <ref>Coulson, p. xxii.</ref>}} is an entirely artificial, literary construct and does not reflect the spoken language.<ref>Coulson, p. xxi.</ref><ref name=Burrow209>Burrow, p. 209.</ref>

==Background== In Sanskrit, as in Proto-Indo-European, a compound is formed by taking the stem-form of the first element (i.e. removing its inflexion){{efn|This process suggests the possibility of an early stage of Proto-Indo-European where words could appear in a sentence without case terminations.}} and combining the two elements with a single accented syllable.

In the later language, this process can be repeated recursively—in theory, ad infinitum, with the freshly made compound becoming the first element of a new one.<ref name=Burrow209/><ref>Whitney, §1246.</ref><ref>Kale, §180</ref>

The process of 'resolving' the compound, i.e., expounding the meaning using the component words declined as in sentence form is termed ''vigraha·vākya''.<ref>Kale, §184.</ref>

Broadly, compounds can be divided into two classes: ''endocentric'' and ''exocentric''.<ref name=exoendo>Adams, p. 35.</ref>

===Endocentric compounds=== An endocentric compound, usually called ''determinative'', is where the compound is essentially the sum of its parts, the meaning being an extension of one of the parts:

* ''blackbird'' → a type of black bird{{efn|a particular species, not just any bird that is black}} * ''White House'' → the official residence of the US president{{efn|the specific presidential residence, not just any house that is white}}<ref>Meyer, p. 179.</ref> * ''siṃha·purás'' → lion-city (Singapore)

===Exocentric compounds=== An exocentric compound refers to something outside the components:

* ''redhead'' → someone with red hair{{efn|that is, not a head}} * ''pickpocket'' → someone who picks someone else's pockets{{efn|that is, not a pocket}}<ref name=exoendo/> * 'bahu·vrīhi' → lit. 'much-rice', i.e., possessing much rice: an indication of wealth

Indeed, this term 'bahuvrihi' is used both in Sanskrit and standard Indo-European linguistics to denote this type of compound.<ref>Fortson, §6.82.</ref>

Sanskrit expands on these to provide several further distinctions as below:

==Classification== In traditional Sanskrit grammar, compounds{{efn|''samāsa·vṛtti''}} are divided into the following main classes:<ref name="Kale, §201">Kale, §201.</ref>

* ''Tatpuruṣa'' ** ''Tatpuruṣa'' proper ** ''Karmadhāraya'' *** ''Dvigu'' ** ''Nañ-tatpuruṣa'' ** ''Prādi'' and ''gati'' ** ''Upapada'' * ''Bahuvrīhi'' ** ''Dvigu'' * ''Dvandva'' * ''Avyayībhāva''

The first two of these, ''tatpuruṣa'' and ''bahuvrīhi'', are Indo-European inheritances, the latter two are Indic innovations.{{Cn|date=December 2023}} Alongside the term ''bahuvrīhi'', ''tatpuruṣa'' has also been adopted in mainstream Indo-European linguistics as the technical term denoting this type of compounding.<ref name=Burrow209/>

The following sections give an outline of the main types of compounds with examples. The examples demonstrate the composition of the compound's elements, and the meanings in English generally correspond to them, in most cases being a similar compound as well. Where this is not the case or the meaning is not clear, a further resolution is provided.

===''Tatpuruṣa'' (determinative)===<!-- Tatpurusha redirects here -->

A ''tatpuruṣa'' is an endocentric compound composed of two elements, wherein the first one, named the attributive, determines the second one.<ref name=Coulson87>Coulson, pp. 87.</ref><ref>Kale, §200.</ref>

Based on the grammatical nature of the attributive member, six varieties of ''tatpuruṣa'' compounds are identified as seen in the classification above. A further distinction is also made based on whether the attributive is in the nominative or an oblique{{efn|accusative, instrumental, dative, ablative, genitive, or locative}} case.<ref name="Kale, §201"/><ref name=Coulson87/>{{efn|The word ''tatpuruṣa'' (a 'that-man', in the sense of 'that person's man', i.e., 'someone's agent') is itself a ''tatpuruṣa'' compound.}}

==== ''Tatpuruṣa proper'' ====

The first member here is an attributive in an oblique relationship with the second, and are therefore termed ''dependent determinatives''.<ref name=Coulson87/><ref>Kale, §203.</ref>

{| class="wikitable" |+ |- ! Word !! Meaning, resolution !! Relationship |- | ''jaya·prepsu''|| victory-seeking || Accusative |- | ''deva·dattá-''|| god-given: given ''by'' the gods || Instrumental |- | ''viṣṇu·bali''|| Viṣṇu-offering: offering ''to'' Viṣṇu || Dative |- | ''svarga·patitá-''|| heaven-fallen: fallen ''from'' heaven || Ablative |- | ''vyāghra·buddhi''|| tiger-thought: the thought ''of'' it being a tiger || Genitive |- | ''yajur·veda-''|| sacrifice-knowledge: the knowledge ''of'' sacrifice || Genitive |- | ''gṛha·jāta-''|| house-born: born ''in'' the house || Locative |}

==== ''Karmadhāraya-tatpuruṣa'' (descriptive) ==== In a ''karmadhāraya-tatpuruṣa'' compound, the first element qualifies the second one adjectively when the latter is a noun. When the second member is an adjective, the qualification is adverbial. Other parts of speech besides adjectives and adverbs may be used to obtain the adjective or adverbial qualification.<ref>Whitney, §1279.</ref><ref>Tubb & Boose, §1.54.</ref>

{| class="wikitable" |+ |- ! Word !! Meaning, resolution |- | ''nīlôtpala''|| blue lotus |- | ''sarva·guṇa''|| all good quality |- | ''priya·sakha''|| dear friend |- | ''maha·rṣí''|| great-sage |- | ''rajata·pātrá''|| silver cup |}

===== ''Dvigu-tatpuruṣa'' (numerative) =====

In essence ''dvigu'' can refer to several compound types where the first element is a numeral.{{efn|The word ''dvi·gu'' itself is an example of dvigu: lit. 'two-cow', used in the sense of 'worth two cows'.}} Dvigu-tatpuruṣa compounds are a special subcategory of ''karmadhārayas''.<ref>Tubb & Boose, §1.56.</ref><ref>Whitney, §1312.</ref>

{| class="wikitable" |+ |- ! Word !! Meaning, resolution |- | ''dvi·rājá-''|| (the battle of) two kings |- | ''tri·bhuvana-''|| three-world: the universe |- | ''tri·yugá-''|| three ages |- | ''tri·divá''|| the triple heaven |- | ''daśâṅgulá-''|| ten fingers' breadth |- | ''ṣaḍ·ahá-''|| six day's time |- | ''sahasrâhṇyá-''|| thousand days' journey |}

''dvigu'' compounds of ''bahuvrīhi'' type are noted below.

==== ''Nañ-tatpuruṣa'' (negative) ==== In a ''nañ-tatpuruṣa'' compound, the first element is a privative, a negator: ''a-'' (before consonant), ''an-'' (before vowel) or ''na-'', just like the English ''un-'', Latin-derived ''in-, non-'' or Greek-derived ''a-'' (before consonant), ''an-'' (before vowel) .<ref>Whitney, §1288.</ref><ref>Tubb & Boose, §1.58.</ref> {| class="wikitable" |+ |- ! Word !! Meaning, resolution |- | ''á·brāhmaṇa''|| non-Brahmin |- | ''án·aśva''|| non-horse: not a horse |- | ''á·pati''|| non-master: not a master |- | ''á·vidyā''|| non-knowledge: ignorance |- | ''á·kumāra''|| unyoung: old |- | ''á·śraddhā''|| unbelief, non-creed |} <!-- Parking this for now, this is a matter of interpretation (both a/an- and na- have already been mentioned above), and may even be splitting hairs. 2021-04-13 While useful in classification, this is said to be etymologically incorrect. Historically, ''a-'' negation originated from ''na'' in its weak grade, Indo-European ''ne'' is normal grade form (Sanskrit: ''na'') and its weak grade is sonant ''ņ'', which in Sanskrit becomes ''a-'' or ''an-'' depending on the following sound.{{Clarify|date=February 2021}} -->

==== ''Upapada-tatpuruṣa'' ====

These are composed of a second member that occurs only in a compound and cannot stand on its own.{{efn|called a bound form}} These are either roots or verbal derivatives from them.<ref>Tubb & Boose, §1.59.</ref><ref>Whitney, §1286.</ref>

{| class="wikitable" |+ |- ! Word !! Meaning, resolution |- | ''sa·yúj''|| joining together |- | ''su·kṛ́t''|| well-doing, a good deed |- | ''kumbha·kāra-''|| pot-maker |- | ''sva·rā́j''|| self-ruling: sovereign |- | ''manu·já''|| Manu-born, born of Manu, man |- | ''svayam·bhū́''|| self-existent |- | ''eka·já''|| only-born |- | ''jala·dá''|| water-giving: a cloud |- | ''sarva·jñá''|| all-knowing: an omniscient person |}

==== ''Aluk-tatpuruṣa'' ==== In an ''aluk-tatpuruṣa'' compound, in contrast to the standard pattern of being in stem form, the first element takes a case form as if in a sentence:<ref>Kale, §208, 215</ref>

{| class="wikitable" |+ |- ! First Element → Word !! Meaning, resolution !! Case |- | ''ojas'' → ''ojasā·kṛtá-''|| done with might <ref>Kale, §206.</ref> || Instrumental |- | ''ātman'' → ''ātmane·pada-''|| 'word for self': the reflexive voice|| Dative |- | ''dyaus'' → ''divas·pati''|| 'lord of the sky': a name of Indra<ref>Kale, §215.</ref><ref>Whitney, §361.</ref> || Genitive |- | ''yudh'' → ''yudhi·ṣṭhira-''|| 'firm in battle': a name<ref>Kale, §217.</ref>|| Locative |}

===''Dvandva'' (co-ordinative)=== These consist of two or more noun stems connected with "and" (copulative or co-ordinative). There are mainly three kinds of ''dvandva'' pair constructions in Sanskrit:<ref>Kale, §187.</ref>

====''Itaretara-dvandva''==== The result of ''itaretara-dvandva''{{efn|enumerative dvandva}} is an enumerative word, the meaning of which refers to all its constituent members. The resultant compound word is in the dual or plural number and takes the gender of the final member in the compound construction. Examples:<ref>Burrow, p. 217.</ref><ref>Whitney, §1253.</ref>

{| class="wikitable" |+ |- ! Word !! Meaning, resolution |- | ''mitrā́·váruṇau''|| Mitra and Varuṇa |- | ''dyāvā·pṛthivī'' || heaven and earth |- | ''vrīhi·yavaú''|| rice and barley |- | ''candrādityau''|| moon and sun |- | ''devāsurā́s''|| the gods and the demons |- | ''hasty·aśvās''|| elephants and horses |- | ''roga·śoka·parītāpa·bandhana·vyasanāni''|| disease, pain, grief, bondage and adversity |}

====''Samāhāra-dvandva''==== Words may be organised in a compound to form a metonym, and sometimes the words may comprise all the constituent parts of the whole. The resultant bears a collective sense and is always singular and neutral.<ref>Kale, §188-189.</ref>

{| class="wikitable" |+ |- ! Word !! Literal ⇒ Meaning |- | ''pāṇi·pādam''|| 'hands and feet' ⇒ limbs, appendages |- | ''āhāra·nidrā·bhayam''|| 'food, sleep and fear' ⇒ vicissitudes, features of life |}

==== ''Ekaśeṣa-dvandva'' ==== Some Sanskrit grammarians identify a third kind of ''dvandva'' which they call ''ekaśeṣa-dvandva'',{{efn|one-(stem)-remains}} where only one stem remains in what is viewed as the compound of multiple words.<ref>Tubb & Boose, §173-174.</ref>

{| class="wikitable" |+ |- ! Compound !! Resolution !! Meaning |- | ''pitarau''|| ''mātā'' + ''pitā''|| 'mother and father' ⇒ parents{{efn|compare Spanish ''padres'' ⇒ 'parents'}} |- | ''mṛgāḥ''|| ''mṛgaḥ'' + ''mṛgā''|| 'does and bucks' ⇒ deer |- | ''putrāḥ''|| ''putrāḥ'' + ''duhitaraḥ''|| 'sons and daughters' ⇒ children |}

==== ''Āmreḍita (iterative)'' ====

While not strictly copulative, this is a compound consisting of the same word repeated with the first occurrence accented.

''Āmreḍita'' compounds are used to express repetitiveness; for example, from ''dív-'' (day) we obtain ''divé-dive'' ('day after day', daily) and from ''devá-'' (god) we obtain ''deváṃ-devam'' or ''devó-devas'' ('deity after deity').<ref>Whitney, §1260.</ref>

===''Bahuvrīhi'' (possessive)=== {{See also|Bahuvrīhi}} ''Bahuvrīhi''{{efn|'much-riced', i.e., a rich person}} is an exocentric compound consisting of a noun preceded by a grammatical modifier which, taken together, functions as a single nominalised adjective.

A ''bahuvrīhi'' compound can often be translated by "possessing..." or "-ed"; for example, "possessing much rice" or "much-riced". In English, examples of ''bahuvrīhi'' would be "lowlife" and "blockhead" (they respectively denote 'one whose life is low' and 'one whose head resembles a block'), or the English surname Longbottom ('one who lives in a long "botham" [valley]').

The second element could essentially have been a noun, which within such a compound, can take on adjective declensions with the compound used adjectivally. Endocentric compounds can thus be transformed into possessives, normally accompanied, and explicitly recognized in the older language, by a change in accentuation:<ref>Whitney, §1293.</ref><ref>Kale, §246.</ref>

* ''indra·śatrú-'', 'Indra's killer' ⇒ ''índra·śatru-'', 'having Indra as killer'{{efn|To illustrate the importance of correct accentuation, the grammarian Patañjali (living around 150 BCE) is said to have cited the story of Vṛtra, who on account of the wrong accent used on this compound ends up instead of being Indra's killer, himself killed by Indra!<ref>Deshpande, p. 24.</ref>}} * ''bṛhad·ratha-'', 'a great chariot' ⇒ ''bṛhád·ratha-'', 'having great chariots' * ''sūrya·tejás-'', 'sun's brightness' ⇒ ''sū́rya·tejas-'', 'possessing the brightness of the sun'

A few typical examples of such compounds:<ref>Burrow, p. 215.</ref>

{| class="wikitable" |+ |- ! Word !! Meaning, resolution |- | ''mayū́ra·roman-''|| peacock-plumed |- | ''ugrá·bāhu-''|| strong-shouldered |- | ''jīvitá·vatsa-''|| alive-childed: having living children{{efn|Indeed in later Sanskrit, the tendency to use compounds instead of verb-based clauses is so strong, in a classical drama, to express gratitude for having children who are alive, one would say the equivalent of 'thank heaven, I'm alive-childed': ''diṣṭyā, jīvita·vats'âsmi'' <ref>Coulson, p. 122.</ref>}} |- | ''mádhu·jihva-''|| honeytongue, honey-tongued: talking nicely |- | ''pátra·hasta-''|| vessel-handed: holding a vessel in the hand |- | ''khara-mukha-''|| donkey-faced |}

==== ''Dvigu-bahuvrīhi'' ==== When the first element of a ''bahuvrīhi'' is a numeral, the compound is called ''dvigu''.{{efn|An example will be the word {{IAST|dvigu}} itself, which as a bahuvrīhi means '[a] two-cow [person]'; i.e., one who has two cows ({{IAST|dvau}} {{IAST|gāvau}}).}} An English example would be a ''halfwit'' ('one who has half of their mind').

A few typical examples of such compounds:<ref>Whitney, §1300.</ref>

{| class="wikitable" |+ |- ! Word !! Meaning, resolution |- | ''éka·cakra-''|| one-wheeled |- | ''éka·pad''|| one-footed |- | ''cátur·aṅga-''|| four-limbed |- | ''saptá·jihva-''|| seven-tongued |- | ''aṣṭá·putra-''|| eight-sonned: having eight sons |- | ''náva·dvāra-''|| nine-doored |- | ''śatá·dant''|| hundred-toothed |- | ''sahásra·nāman''|| thousandname: having a thousand names |}

===''Avyayībhāva'' (adverbial)=== ''Avyayībhāvas'' ('indeclinable') are adverbial compounds composed of an indeclinable element (an adverb, etc.) and a noun, together expressing an adverb or another indeclinable ({{IAST|avyaya}}) element.<ref>Whitney, §1313.</ref><ref>Tubb & Boose, §1.48.</ref>

{| class="wikitable" |+ |- ! Word !! Meaning, resolution |- | ''adhy·ātmam''|| concerning the self |- | ''anu·ratham''|| behind the chariots |- | ''upa·rājam''|| near the king; kingside |- | ''praty·agni''|| towards the fire |- | ''prati·niśam''|| every night |- | ''yathā·śakti''|| per-strength: according to one's ability |- | ''sa·cakram''|| simultaneously with the wheel |- | ''antar·jalam''|| inter-water: within the water |} <!-- Park this for now, not attested by any of the noted reliable sources. ===''Āmreḍita'' (iterative)=== A compound consisting of the same word repeated twice or more but with the first occurrence being accented.<ref>http://findarticles.com/p/articles/mi_go2081/is_200310/ai_n9761222 {{Bare URL inline|date=May 2022}}</ref> ''Amreditas'' are used to express repetitiveness; for example, from ''dív'' (''day'') we obtain ''divé-dive'' ('day after day', daily) and from ''devá'' (god) we obtain ''{{IAST|deváṃ-devam}}'' or ''devó-devas'' ('deity after deity').<ref>http://www.utexas.edu/cola/centers/lrc/eieol/vedol-EI-X.html {{Dead link|date=February 2022}}</ref> -->

==See also== *Sanskrit nominals *Sanskrit verbs *Sanskrit grammar *Vedic Sanskrit grammar *Proto-Indo-European verbs *Proto-Indo-Aryan *Proto-Indo-Iranian *Proto-Indo-European *Kenning

== Notes == {{notelist|20em}}

==References== {{reflist|15em}}

==Bibliography== * {{cite book |last1=Coulson |first1=Michael |title=Sanskrit (Teach yourself) |date=2003 |publisher=Oxford |location=Great Britain |isbn=0-340-32389-2 |edition=2003}} * {{cite book |last1=Fortson |first1=Benjamin W |title=Indo-European Language and Culture |year=2010 |publisher=Wiley-Blackwell |isbn=978-1-4051-8895-1 |edition=2010}} * {{cite book |last1=Burrow |first1=T |title=The Sanskrit Language |year=2001 |publisher=Motilal Banarsidass |isbn=81-208-1767-2 |edition=2001}} * {{cite book |last1=Whitney |first1=William Dwight |title=Sanskrit Grammar |date=January 2008 |publisher=Motilal Banarsidass |isbn=978-81-208-0620-7 |edition=2000}} * {{cite book |last1=Macdonnel |first1=Arthur Anthony |title=A Sanskrit Grammar for Students |year=1997 |publisher=Motilal Banarsidass |isbn=81-246-0094-5}} * {{cite book |last1=Macdonnel |first1=Arthur Anthony |title=A Vedic Sanskrit Grammar for Students |year=1993 |publisher=Motilal Banarsidass |isbn=81-208-1053-8 |edition=2000}} * {{cite book |last1=Goldman |first1=Robert P |title=Deva·vāṇī́·praveśikā|year=2019 |publisher=Center for South Asia Studies, University of California |isbn=978-0-944613-40-5}} * {{cite book |last1=Kale |first1=M R |title=A Higher Sanskrit Grammar |year=1969 |publisher=Motilal Banarsidass |isbn=81-208-0177-6 |edition=2002}} * {{cite book |last1=Adams |first1=Valerie |title=An Introduction to Modern English Word-Formation |date=1987 |publisher=Longman Group |isbn=0-582-55042-4}} * {{cite book |last1=Meyer |first1=Charles |title=Introducing English Linguistics |date=2009 |publisher=Cambridge University Press |isbn=978-0-521-83350-9 |edition=1st}} * {{cite book |last1=Carstairs-McCarthy |first1=Andrew |title=An Introduction to English Morphology |date=2002 |publisher=Edinburgh University Press |isbn=0-7486-1326-9}} * {{cite book |last1=Tubb|last2=Boose |first1=Gary|first2=Emery |title=Scholastic Sanskrit |date=2007 |publisher=American Inst of Buddhist Studies |location=NY |isbn=978-0-9753734-7-7}} * {{cite book |last1=Deshpande |first1=Madhav M |title=Sanskrit and Prakrit |year=1993 |publisher=Motilal Banarsidass |isbn=81-208-1136-4 |edition=1993}}

compound Category:Vyakarana Category:Linguistic morphology