Typology and Universals: Structural Types, Greenberg, South Asian Families and Language Law

This chapter answers Section XH3.5 of XH3 Linguistics, "Typology and Universals", one of the six optional sections of the Humanities and Social Sciences (XH) paper, chosen at application. The heading names the classification of languages by structural type (agglutinative, isolating, fusional) and by family tree (genetic classification), language universals including Greenberg's word-order universals, the major South Asian language families (Indo-Aryan, Dravidian, Austroasiatic, Tibeto-Burman) with their shared areal features, and the constitutional provisions for languages in India. The chapter separates three ways of grouping languages, works through implicational universals, places the families of South Asia and the convergence that makes the region a linguistic area, and sets out the Eighth Schedule and the official-language articles of the Constitution.

1. Structural (morphological) types of language

Typology groups languages by shared structure, regardless of their family. The classic morphological typology sorts languages by how many morphemes they pack into a word and how cleanly those morphemes can be separated. The scale runs from isolating (close to one morpheme per word) to highly synthetic (many per word), and these are tendencies, not boxes: most languages mix types and sit somewhere on a continuum measured by an index of synthesis (the average number of morphemes per word).

The main morphological types
TypeHallmarkExamples
Isolating (analytic)About one morpheme per word; grammar carried by word order and separate particlesMandarin Chinese, Vietnamese
AgglutinativeSeveral affixes strung in a row, each with one clear meaning and a clear boundaryTurkish, Tamil, Finnish, Swahili
Fusional (inflectional)One affix fuses several categories at once, with no clean cut between themLatin, Sanskrit, Russian
PolysyntheticVery many morphemes per word, often incorporating the object, so a word can be a sentenceInuktitut, Mohawk
⚠️ Agglutinative versus fusional
Both pack many morphemes into a word, so the test is separability. In an agglutinative word each affix maps to one meaning and you can cut cleanly between them; in a fusional word one ending carries several meanings fused together (in Latin domin-us, the ending -us marks nominative and singular at once) and cannot be cut apart. Counting morphemes alone will not tell the two apart.

2. Genetic classification and its rivals

Languages can be grouped in three quite different ways, and keeping them apart is a frequent exam point. Genetic (genealogical) classification groups languages by common descent from a single ancestral proto-language, drawn as a family tree; membership is permanent and is established by the comparative method. Typological classification groups by shared structure (the morphological types above, or word order), and a language can change type over time. Areal classification groups by features shared through contact in a region. The three cut across one another: English and Hindi are genetically related (both Indo-European) yet typologically different, while Hindi and Tamil are genetically unrelated yet areally alike.

🎯 Why type is not descent
Two languages can be the same type (both agglutinative, say) by accident or by contact, without any shared ancestor, and two languages from one family can drift into different types. So a structural resemblance never proves relatedness; only regular sound correspondences, worked out by the comparative method, establish a family.

3. Language universals and Greenberg's word-order universals

A language universal is a property true of all or most languages. Universals come in two cross-cutting pairs. An absolute universal holds without exception; a tendency holds of most languages. An unrestricted universal states a fact outright (all languages have vowels); an implicational universal has the form 'if a language has X, then it also has Y'. Implicational universals are powerful because one visible property lets you predict another.

Joseph Greenberg (1963), from a 30-language sample, proposed a set of word-order universals built on the basic order of subject (S), object (O) and verb (V). The commonest basic orders worldwide are SOV (Hindi, Tamil, Japanese, Turkish) and SVO (English, Mandarin), followed by VSO (Classical Arabic, Irish); the object-first orders are rare. His Universal 1 notes that the subject almost always precedes the object. Universal 3 says VSO languages are prepositional, and Universal 4 that SOV languages are overwhelmingly postpositional. These feed a broader pattern of harmony: verb-object (VO) languages tend to use prepositions and to put the noun before its genitive, while object-verb (OV) languages tend to use postpositions and to put the genitive before the noun.

Other Greenberg universals follow the same implicational logic. Universal 2: in languages with prepositions the genitive (possessor) almost always follows the noun it modifies, while in languages with postpositions it nearly always precedes it. Universal 34: no language has a trial number unless it has a dual, and none has a dual unless it has a plural. Universal 25: if the pronominal object follows the verb, so does the nominal object. Predictions can be chained. A language described only as verb-final (SOV) is expected, by Universal 4, to be postpositional and then, by Universal 2, to put the possessor before the possessed noun, as Hindi does in 'Ram ka ghar' (Ram's house). Each universal lets you derive something unseen from something seen, and each is a strong tendency, not an exceptionless law.

ℹ️ Tendency, not law
Most of Greenberg's word-order universals are statistical tendencies ('with overwhelmingly more than chance frequency'), not exceptionless laws. A single language that is SOV yet prepositional does not refute Universal 4; it is a rare counter-case to a strong tendency. Treating a tendency as absolute is the usual error.

4. The language families of South Asia

South Asia is home to four major language families, plus smaller groups and isolates. Distinguishing them by membership is a reliable exam target.

  • Indo-Aryan, a branch of Indo-European, covers most of the north and centre: Hindi, Urdu, Bengali, Marathi, Gujarati, Punjabi, Odia, Assamese, Sindhi, Nepali and Sinhala descend through the Prakrits from Old Indo-Aryan.
  • Dravidian, concentrated in the south, has four major literary languages — Tamil, Telugu, Kannada and Malayalam — plus Tulu, Gondi, Kurukh and, as a far outlier, Brahui in Pakistan's Balochistan. These languages are agglutinative and rich in retroflex consonants.
  • Austroasiatic is represented in India by the Munda languages (Santali, Mundari, Ho) of the east and centre, and by Khasi in Meghalaya and the Nicobarese languages.
  • Tibeto-Burman, a branch of Sino-Tibetan, is spoken across the north-east and the Himalayas: Bodo, Manipuri (Meitei), Mizo, the Naga languages and many more. Smaller groups (Tai, Andamanese) and isolates also occur.
Wider affiliation and typical profile of the four families
FamilyWider affiliationTypical profile in South Asia
Indo-AryanIndo-EuropeanVerb-final order, postpositions, case and gender marking
DravidianA separate family; no outside link is firmly establishedAgglutinative suffixes, verb-final order, postpositions
AustroasiaticA large family reaching into mainland Southeast AsiaMunda branch largely suffixing and verb-final
Tibeto-BurmanSino-TibetanMostly verb-final; many languages are tonal

5. South Asia as a linguistic area

Despite belonging to four unrelated families, the languages of South Asia share a striking set of features, acquired through long contact rather than inherited from a common ancestor. This convergence makes the region a linguistic area (Sprachbund), a point argued in Murray Emeneau's classic 1956 study 'India as a Linguistic Area'. A feature that recurs across Indo-Aryan, Dravidian and Munda alike is a marker of the area, not of a family.

  • Retroflex consonants (a dental versus retroflex contrast) run through the Indo-Aryan, Dravidian and Munda families and are the best-known areal feature.
  • Verb-final (SOV) order with postpositions across most of the families, and the dative (experiencer) subject construction, in which the one who feels or knows is marked like a recipient rather than an agent.
  • Conjunctive participles (converbs, 'having done X, ...'), compound (vector) verbs, echo-word reduplication, and quotative constructions recur across the families.

6. Constitutional provisions for languages in India

The Constitution of India handles the country's many languages through several provisions. Article 343 makes Hindi in the Devanagari script the official language of the Union, with English continued for official purposes. Article 344 provides for an official-language commission and committee. Articles 345 to 347 deal with the official language or languages of the states, the language for communication between states, and recognition of a language spoken by a section of a state's population. Article 348 concerns the language of the higher courts. Article 350A provides for mother-tongue instruction at the primary stage, and Article 350B for a Special Officer for Linguistic Minorities. Article 351 directs the Union to develop Hindi.

The Eighth Schedule lists the scheduled languages that enjoy special constitutional recognition; it began with 14 languages and now contains 22 after successive amendments added Sindhi, then Konkani, Manipuri and Nepali, and later Bodo, Dogri, Maithili and Santali. Its languages are the ones referred to in Articles 344(1) and 351. A point the exam likes to test is that the Constitution names Hindi an official language of the Union, not a 'national language': India has official languages, not a constitutionally declared national language.

Four more points complete the picture. The Official Languages Act, 1963 allowed English to continue in official use for the Union beyond 1965, alongside Hindi. Article 120 makes Hindi or English the language of business in Parliament, with the presiding officer able to permit a member to speak in the mother tongue. Article 29(1) gives any section of citizens with a distinct language, script or culture the right to conserve it, and Article 30(1) gives religious and linguistic minorities the right to establish and administer educational institutions of their choice. Article 347 lets the President direct that a language spoken by a substantial proportion of a state's population be officially recognised in that state, on demand.

⚠️ Official is not national
Calling Hindi the 'national language' is a common mistake. The Constitution makes Hindi (with English) an official language of the Union for government business; it does not declare any national language. Equally, the Eighth Schedule's 22 languages are scheduled languages, which is not the same as being official languages of the Union.

Key takeaways

  • Morphological typology ranks languages by morphemes per word: isolating (about one), agglutinative (many, cleanly separable, one meaning each), fusional (one affix fuses several categories) and polysynthetic (very many, with incorporation). These are continua, and separability distinguishes agglutinative from fusional.
  • Three classifications must be kept apart: genetic (common ancestor, permanent, by the comparative method), typological (shared structure, can change) and areal (shared by contact). Structural likeness never proves relatedness; only regular sound correspondences do.
  • Universals are absolute or tendencies, and unrestricted or implicational ('if X then Y'). Greenberg (1963) found the subject usually precedes the object (U1); VSO languages are prepositional (U3) and SOV languages overwhelmingly postpositional (U4); VO and OV orders show harmonic correlations. Most are tendencies, not absolutes.
  • South Asia has four major families: Indo-Aryan (Indo-European branch; Hindi, Bengali, Marathi, etc.), Dravidian (Tamil, Telugu, Kannada, Malayalam; Brahui as a far outlier), Austroasiatic (Munda languages, Khasi) and Tibeto-Burman (Sino-Tibetan branch; Bodo, Meitei, Mizo, Naga).
  • Long contact has made South Asia a linguistic area (Emeneau, 1956): retroflex consonants, SOV order with postpositions, dative-subject constructions, conjunctive participles, compound verbs, echo words and quotatives recur across unrelated families, a convergence of area rather than descent.
  • The Constitution makes Hindi in Devanagari the official language of the Union (Article 343), with English continued; Articles 350A/350B protect mother-tongue instruction and linguistic minorities. The Eighth Schedule now lists 22 scheduled languages (from an original 14). Hindi is an official, not a national, language.

Practice questions (18)

Attempt each one before opening the answer. Every explanation names the tempting wrong option as well as the right one, because that is where marks are lost.

  1. A language in which words are built by stringing together several affixes, each carrying exactly one grammatical meaning and separable by a clean boundary, is classified as:

    1. Isolating, because each separate word tends to consist of just one morpheme with no affixes
    2. Polysynthetic, because the object is incorporated so that a single word can express a whole sentence
    3. Fusional, because a single ending fuses several grammatical categories together at the same time
    4. Agglutinative, because affixes are glued in a row with one meaning each and clear cuts
    Show answer

    Answer: D — Agglutinative, because affixes are glued in a row with one meaning each and clear cuts

    One-meaning-per-affix with clean boundaries is the hallmark of an agglutinative language (Turkish, Tamil). Isolating languages have about one morpheme per word; fusional languages fuse several categories into one unanalysable ending; polysynthetic languages pack very many morphemes with incorporation. The clean separability is what picks out agglutinative.

  2. Mandarin Chinese expresses most grammatical relations by word order and separate particles rather than by inflecting its words, which carry little or no inflectional morphology. Mandarin is a good example of which type?

    1. Isolating, with grammar carried by order and particles rather than by inflection
    2. Fusional, with single endings that fuse case, number and gender all together, as in Latin or Russian nouns
    3. Polysynthetic, with many morphemes and the object incorporated into the verb form, as in Inuktitut
    4. Agglutinative, with long chains of clearly separable one-meaning affixes, as in Turkish or Tamil
    Show answer

    Answer: A — Isolating, with grammar carried by order and particles rather than by inflection

    A language with little or no inflection that relies on order and particles is isolating (analytic), and Mandarin is the standard example. Fusional, polysynthetic and agglutinative types all build words by heavy inflection, incorporation or chains of affixes, which is exactly what Mandarin largely avoids, so none of them fits.

  3. Which of the following statements about morphological typology are correct?

    1. In a fusional language a single affix may express several grammatical categories at once
    2. Morphological types are tendencies on a continuum, and a single language may mix them
    3. The morphological type of a language is fixed forever and can never change over time
    4. In an agglutinative language affixes are typically separable, each mapping to one meaning
    Show answer

    Answer: A — In a fusional language a single affix may express several grammatical categories at once; B — Morphological types are tendencies on a continuum, and a single language may mix them; D — In an agglutinative language affixes are typically separable, each mapping to one meaning

    Fusional affixes fuse categories; types are continua that a language can mix; and agglutinative affixes are separable with one meaning each. The claim that a type is fixed forever is false: languages drift between types over time (for example, from more synthetic toward more analytic), as the history of English shows.

  4. Genetic (genealogical) classification groups languages on the basis of:

    1. The number of morphemes they typically combine within a single word
    2. The geographical region in which their speakers currently happen to live
    3. Their descent from a common ancestral proto-language
    4. The basic order of subject, object and verb that they employ
    Show answer

    Answer: C — Their descent from a common ancestral proto-language

    Genetic classification is based on descent from a shared proto-language, established by the comparative method. Morpheme count and word order are typological criteria, and geographical region is the basis of areal classification, none of which is about common ancestry.

  5. Hindi and Tamil share verb-final order, postpositions and retroflex consonants, yet belong to different families, while Hindi and English belong to one family but differ sharply in structure. What does this pattern show?

    1. That structural similarity reliably proves that two languages must therefore be genetically related
    2. That English and Hindi cannot really be related because their structures are so strikingly different
    3. That Hindi and Tamil must descend from one proto-language because they look so much alike in structure
    4. That genetic and typological classifications are independent and can cut across each other
    Show answer

    Answer: D — That genetic and typological classifications are independent and can cut across each other

    The pattern shows that genetic relatedness and structural type are independent: related languages can look different and unrelated ones alike. So structural similarity does not prove relatedness (Hindi and Tamil are not thereby one family), and structural difference does not disprove it (Hindi and English remain Indo-European); the shared structure of Hindi and Tamil is areal, not genetic.

  6. A universal of the form 'if a language has property X, then it also has property Y' is classed, by its logical form, as:

    1. An implicational universal, since one property implies the presence of another
    2. An unrestricted universal, since it states a fact true of all languages outright
    3. A genetic marker, since it identifies membership of one particular language family
    4. A typological index, since it measures the morphemes contained per word
    Show answer

    Answer: A — An implicational universal, since one property implies the presence of another

    A conditional of the form 'if X then Y' is an implicational universal by its logical form. An unrestricted universal asserts a property outright, with no 'if'; a genetic marker identifies family membership rather than the form of a universal; and a typological index is a morpheme-count measure, not a universal at all. Whether an implicational universal is exceptionless or only a tendency is a separate question.

  7. By Greenberg's Universal 3, a language whose dominant basic order is VSO is predicted to be:

    1. Postpositional, placing its adpositions after the noun they go with
    2. Prepositional, placing its adpositions before the noun they go with
    3. Without any adpositions at all, marking relations only by case endings
    4. Free in adposition order, using prepositions and postpositions equally often
    Show answer

    Answer: B — Prepositional, placing its adpositions before the noun they go with

    Universal 3 states that VSO languages are prepositional, part of the harmony linking verb-object order to prepositions. Postpositions pattern with SOV order (Universal 4), not VSO; and the claims that a VSO language has no adpositions or uses both equally are not what Greenberg's generalization predicts.

  8. Consider the six logically possible basic orderings of subject (S), object (O) and verb (V): SOV, SVO, VSO, VOS, OVS, OSV. Greenberg's Universal 1 observes that the dominant order almost always places the subject before the object. Of these six orderings, how many place S before O? Give the number only.

    Numerical answer — type the value.

    Show answer

    Answer: 3

    Check each order for S before O: SOV (S before O, yes), SVO (yes), VSO (yes), VOS (object before subject, no), OVS (no), OSV (no). Three of the six place the subject before the object, and these three (SOV, SVO, VSO) are exactly the common orders Universal 1 describes.

  9. Greenberg's implicational generalizations say that SOV languages are overwhelmingly postpositional (Universal 4) and VSO languages are prepositional (Universal 3). Six languages are described: Alpha is SOV with postpositions; Beta is SOV with postpositions; Gamma is VSO with prepositions; Delta is SOV with postpositions; Epsilon is SOV with prepositions; Zeta is VSO with postpositions. How many of the six conform to these two implicational generalizations? Give the number only.

    Numerical answer — type the value.

    Show answer

    Answer: 4

    Alpha, Beta and Delta are SOV with postpositions, which conforms to Universal 4; Gamma is VSO with prepositions, which conforms to Universal 3 — four conform in all. Epsilon is SOV yet prepositional (against Universal 4) and Zeta is VSO yet postpositional (against Universal 3), so these two go against the generalizations. The count of conforming languages is four.

  10. Which of the following statements about language universals and word order are correct?

    1. SOV and SVO are the two commonest basic word orders among the world's languages
    2. An absolute universal holds without exception, whereas a tendency holds of most languages
    3. Most of Greenberg's word-order universals are statistical tendencies rather than exceptionless laws
    4. Object-initial orders such as OSV and OVS are the most frequent orders in the world
    Show answer

    Answer: A — SOV and SVO are the two commonest basic word orders among the world's languages; B — An absolute universal holds without exception, whereas a tendency holds of most languages; C — Most of Greenberg's word-order universals are statistical tendencies rather than exceptionless laws

    SOV and SVO are indeed the two commonest orders; the absolute-versus-tendency distinction is correctly stated; and most of Greenberg's order universals are tendencies. The last claim is false: object-initial orders (OSV, OVS) are among the rarest, not the most frequent, which is why the common orders all keep the subject before the object.

  11. Which set lists only languages that belong to the Dravidian family?

    1. Marathi, Gujarati and Odia, the major languages of the western and eastern coastal states
    2. Santali, Mundari and Ho, the Munda languages spoken across eastern and central India
    3. Tamil, Telugu, Kannada and Malayalam, the four major literary languages of the south
    4. Bodo, Meitei and Mizo, languages spoken across the north-east and the eastern Himalayan region
    Show answer

    Answer: C — Tamil, Telugu, Kannada and Malayalam, the four major literary languages of the south

    Tamil, Telugu, Kannada and Malayalam are the four major Dravidian literary languages. Marathi, Gujarati and Odia are Indo-Aryan; Santali, Mundari and Ho are Austroasiatic (Munda); and Bodo, Meitei and Mizo are Tibeto-Burman, so none of those sets is Dravidian.

  12. To which language family do the Munda languages such as Santali, Mundari and Ho belong?

    1. Tibeto-Burman, the branch of the larger Sino-Tibetan family spoken across India's north-east
    2. Austroasiatic, which in India also includes Khasi and the Nicobarese languages
    3. Indo-Aryan, the branch of Indo-European that covers the northern and central plains of the subcontinent
    4. Dravidian, the southern family that also includes Tamil, Telugu and Kannada
    Show answer

    Answer: B — Austroasiatic, which in India also includes Khasi and the Nicobarese languages

    The Munda languages are the main Indian branch of Austroasiatic, which also includes Khasi and the Nicobarese languages. They are not Tibeto-Burman, Indo-Aryan or Dravidian; those are the other three major South Asian families, each with its own distinct membership.

  13. Which of the following statements about South Asian language families are correct?

    1. Indo-Aryan is a branch of the Indo-European family and includes Hindi and Bengali
    2. Brahui, spoken in Pakistan's Balochistan, is a far north-western outlier of Dravidian
    3. Tibeto-Burman languages are a branch of the Sino-Tibetan family
    4. Tamil and Telugu are Indo-Aryan languages descended from the Prakrits
    Show answer

    Answer: A — Indo-Aryan is a branch of the Indo-European family and includes Hindi and Bengali; B — Brahui, spoken in Pakistan's Balochistan, is a far north-western outlier of Dravidian; C — Tibeto-Burman languages are a branch of the Sino-Tibetan family

    Indo-Aryan is an Indo-European branch (Hindi, Bengali); Brahui is a Dravidian outlier in Balochistan; and Tibeto-Burman is a Sino-Tibetan branch. The last claim is false: Tamil and Telugu are Dravidian, not Indo-Aryan, and do not descend from the Prakrits, which are the ancestors of the Indo-Aryan languages.

  14. Retroflex consonants are found across the Indo-Aryan, Dravidian and Munda languages of the subcontinent, even though these belong to unrelated families. In the study of South Asia, this shared trait is best described as:

    1. An areal feature, shared by contact across families within a linguistic area
    2. A genetic feature, proving that the three families share one common ancestor
    3. A morphological type, placing all three families in the agglutinative class
    4. A pure coincidence, unconnected with any history of contact among the languages
    Show answer

    Answer: A — An areal feature, shared by contact across families within a linguistic area

    A trait shared across unrelated families through long contact is an areal feature, the mark of a linguistic area (Emeneau's South Asia). It cannot be a genetic feature, since the families are unrelated; it is not a morphological type; and its systematic spread across the region is contact-driven, not coincidence.

  15. How many languages are currently listed in the Eighth Schedule of the Constitution of India?

    1. 18 languages, the number that applied before the most recent amendments were passed
    2. 22 languages, after amendments added Sindhi, Konkani, Manipuri, Nepali and four more
    3. 14 languages, the original number included when the Constitution first came into force
    4. 25 languages, including every language with official status in any state of the Union
    Show answer

    Answer: B — 22 languages, after amendments added Sindhi, Konkani, Manipuri, Nepali and four more

    The Eighth Schedule currently lists 22 languages: it began with 14 and grew by amendment (Sindhi; then Konkani, Manipuri and Nepali; then Bodo, Dogri, Maithili and Santali). 14 was the original figure, 18 was an intermediate figure, and there is no 25-language list, so the other counts are wrong.

  16. A candidate claims that 'Hindi is the national language of India under the Constitution.' Why is this statement inaccurate?

    1. Because the Constitution names no official language at all and leaves the matter to the states
    2. Because the Constitution makes English, not Hindi, the sole official language of the Union
    3. Because Article 343 makes Hindi an official language, and no national language is named
    4. Because Hindi appears in the Eighth Schedule, which automatically bars it from any official role
    Show answer

    Answer: C — Because Article 343 makes Hindi an official language, and no national language is named

    Article 343 makes Hindi in Devanagari an official language of the Union (with English continued), but the Constitution declares no national language, so calling Hindi the 'national language' is inaccurate. It is wrong that the Constitution names no official language, wrong that English alone is official, and wrong that listing in the Eighth Schedule bars a language from official use.

  17. A newly described language has SOV basic order, and nothing else about it is known. Apply two tendencies in sequence. Universal 4: SOV languages are overwhelmingly postpositional. Universal 2: prepositional languages nearly always put the genitive (possessor) after the noun, and postpositional languages nearly always put it before the noun. Which pair of predictions follows?

    1. It is prepositional, and it places the possessor after the possessed noun
    2. It is postpositional, and it places the possessor after the possessed noun
    3. It is prepositional, and it places the possessor before the possessed noun
    4. It is postpositional, and it places the possessor before the possessed noun
    Show answer

    Answer: D — It is postpositional, and it places the possessor before the possessed noun

    Universal 4 takes SOV to postpositions, and Universal 2 takes postpositions to a possessor placed before the noun, so the chain gives postpositional with the possessor first. The other pairs each break one link: a prepositional prediction contradicts Universal 4, and a possessor after the noun is what Universal 2 predicts only for prepositional languages, not for the postpositional language reached at the first step. Both are tendencies, so the chain is a prediction, not a proof.

  18. Which provision of the Constitution of India addresses facilities for instruction in the mother tongue at the primary stage of education for children of linguistic minority groups?

    1. Article 350A, which asks states and local authorities to provide primary-stage mother-tongue instruction
    2. Article 343, which makes Hindi in Devanagari the official language of the Union and continues English
    3. Article 351, which directs the Union to promote the spread and the development of the Hindi language
    4. Article 120, which makes Hindi or English the language of business in the houses of Parliament
    Show answer

    Answer: A — Article 350A, which asks states and local authorities to provide primary-stage mother-tongue instruction

    Article 350A deals with mother-tongue instruction at the primary stage for linguistic minority children; it is paired with Article 350B, which provides for a Special Officer for Linguistic Minorities. Article 343 fixes the official language of the Union, Article 351 directs the development of Hindi, and Article 120 governs the language of business in Parliament, so none of those concerns primary-stage mother-tongue teaching.