Typology (by Edward J. Vajda)

Differences between the use of free and bound morphemes in the world's languages.

      In English, many root morphemes are free.  And many English content and function words consist of single morphemes: hunt, kill, the.  The same is even more true of Chinese, Vietnamese and a few other East Asian languages.  This is not the case in all languages of the world.  In many languages, bound roots are the rule rather than exception. Let's look at data from three languages where words tend to be morphologically complex and roots tend to be bound: Navaho and Swahili.

      The morphology of the Navaho verb is striking in its lack of free root morphemes:  all Navaho verb roots without exception are bound morphemes.  For example, the Navaho morpheme that conveys the basic meaning of the English give is ní.  When speaking Navaho, one cannot say anymore than an English-speaking person can say -ject- or -mit- and expect to be understood.  Besides its basic, concrete meaning, each Navaho verb form conveys not only a basic idea such as give, take, have; it must also contain morphemes conveying tense, number of subjects, duration or repetition of action and even the shape of the object for those verbs that take direct objects. 

      The imperative, or command for, is morphologically the most simple verb stem.  And yet there are over 20 separate verb suffixes conveying the shape and type of a verb's object.  We said that ní means Give, but even in the meaning of a basic command to hand the speaker an object it can never be used as a separate word unless it is attached to a suffix that marks the shape of the object to be given:

Awéé' shaa níLteeh. Give me the baby.  (-Lteeh = living)
TL'ooL shaa ní. Give me the rope. (-lé = long flexible)
Tó shaa níkaah. Give me the water. (-kaah = liquid in a container)
Tsin shaa níti,i,h. Give me the pole. (-ti,i,h= slender, rigid)
Beeldléí shaa níLtsóós. Give me the blanket.  (-Ltsóós = flat)
Atsi' shaa ní'aah. Give me the meat. (-'aah = dense bulky object)

Navaho has about 20 such verbal object markers altogether.  The same noun might be used in conjunction with more than one verbal marker.

- Naaltsoos, book  or paper 

Naaltsoos shaa níLtsóós. Give me the paper.  (-Ltsóós flat) 

Naaltsoos shaa ní'aah. Give me the book.  (-'aah = dense bulky object) 

- tL'iish, snake

TL'iish shaa níLteeh.       Give me the snake.  (living) 

TL'iish shaa ní.              Give me the dead snake.  (-lé = long flexible) 

      In Navaho only a very few root morphemes denoting basic concrete concepts can stand as separate words: shash bear, chizh firewood, oljéé moon.  All of the remaining words in Navaho contain special affixes. Thus, most noun roots denoting basic concepts are bound morphemes.

      Noun morphemes that denote body parts, family relations, or important objects everyone is expected to possess such as shoes, clothing, etc. cannot be used as separate words but must always be accompanied by a prefix denoting the possessor.  Thus, amazingly enough, there is no independent Navaho word to mean mother or father, sister or brother, ear, tooth, or leg, or even shoes or bow and arrow.  Navaho can express all these concepts, of course, but not on the level of the simple word.  In Navaho, one must use morphologically complex words meaning someone's mother, someone's father: For instance: shizhé'é my father, nizhé'é your father; bizhé'é his/her father ; or--at best-- 'azhé'é which has the general meaning of one's, some undetermined person's father.  Conversely, if I want to assert possession over such a person or object that intrinsically is not by own, I say literally my someone else's:  she'e' 'awoo' my someone else's toothShiwoo' means my tooth in the sense of a part of my body.

      Because nearly all Navaho words have a complex internal morphology, Navaho borrows very few words from other languages.  Perhaps as few as 30 nouns in Navaho are known to be borrowed, no verbs:  mosi, chidí, béeso.  In English, 75% of the words are borrowed.  So the morphological structure of a particular language may influence why one language borrows many words and another does not.

      Let's turn to an even more extreme case. In Swahili all noun roots, even ones denoting the most basic concepts, are bound morphemes which must be accompanied by a prefix marking which word class the noun belongs to:

Human class: (m/wa)           mtu  (person),          mtoto (child),                 mgeni (stranger) 

Things class: (ki/vi)              kisu (knife),             kikapu (basket),            kitabu (book).

The same is true of adjectives and number roots; they require the same prefixes depending upon the noun they are attached to:

numbers:        -moja (1), --wili (2), -tatu (3), -mne (4), -tano (5)

adjectives: -dogo (small), -baya (bad), -zuri (good)

Language Typology--favored technique of word building.

It is possible to categorize entire languages according to how words are built up.  There are several types. 

1) Languages that tend not to combine morphemes at all exept to form compounds are called isolating languages.  Chinese and Vietnamese are the most purely isolating, with nearly half their words being monomorphemic and also monosyllabic.  Since analytic languages usually have relatively few derivational or inflectional morphemes, they tend to form words by combining free morphemes into compounds.  Thus a preponderance of both simple and compound words tend to be a feature of analytic languages.  Nearly half of all Chinese words are monomorphic, and most of the rest are compounds of the type gascan, blackboard, etc, such as dianhua telephone; diannao computer.

2) agglutination is the addition of a large number of affixes one after another: book-s, re-use, lik-able, anti-dis-establish-ment-ari-an-ism.  (Swahili and other Bantu languages)  An agglutinating language is one whose primary means of building new words is by adding affixes.  There are agglutinating languages where prefixation predominates (Quileute); others prefer suffixation (Hungarian haz-ak-ban in the houses; Turkic languages: Kazakh it-ter-in of the dogs). In Australia there is a family of languages that is exclusively suffixing and another that is exclusively prefixing.

3) fusion: type of agglutination where each morpheme changes the shape of the ones to which it is added:  long--length.  (Russian, Latin, Cherokee)  An inflecting language is one whose primary means of building new words is by adding affixes, and adjacent morphemes tend to cause phonological changes in one another.  Different positional variants of the same morpheme are called allomorphs.  Sometimes agglutinating languages are also referred to as inflecting languages because they also make extensive use of grammatical affixes.

4) symbolic fusion: internal changes replace inflections (English:  foot--feet; tooth--teeth).  Hebrew, Arabic, and other Afroasiatic languages make the most use of this technique (Arabic root s/l/m).  Other languages make sporadic use: Tibetan g-tong-ba to give, b-tang gave; thong Give!  Even English make some use of this technique:  tooth--teeth, foot-feet, run, ran, but, historically, these exceptions are remnants of regular fusion. 

      Notice the difference between infixing and symbolic fusion.  In English f+vowel+t is the morpheme.  Whereas in an infixing language the infix doesn't replace any part of the root morpheme into which it is added:  Bontoc (Phillipines) fikas strong--fUMikas to be strong

      Once again, the difference in technique is one of degree, since the same language can use all the techniques to a greater or lesser degree.  Finnish or Turkish are called agglutinating languages because they favor this technique of word building over others. English is called a fusional language because if favors the technique of fusion.  But it also uses agglutination, and even symbolic fusion: run, ran.

      No language uses reduplication, blending, or other minor word formation types as its favored technique of word building.

Language typology--how much grammatical information is stored in the word? 

      Let us now turn to how the structure of words and morphemes may be used to categorize entire languages.  Comparing various languages according to structure is called language typology.  Typology may involve any structural aspect of language:  phonology morphology or syntax.  Let's look at how languages can be classified according to the number and complexity of morphemes used to build words.

Expression of grammatical meaning in the word and the sentence

Give example sentences:

      In analytic languages words tend to consist of free morphemes, i.e., they are monomorphic.  Chinese and Viet. are the most purely analytical:  nearly half their words are monomorphic, and also monosyllabic.  Since analytic languages usually have relatively few derivational or inflectional morphemes, they tend to form words by combining free morphemes into compounds.  Thus a preponderance of both simple and compound words tend to be a feature of analytic languages.  Nearly half of all Chinese words are monomorphemic, and most of the rest are compounds of the type gascan, blackboard, etc, such as dianhua telephone, or diankan TV.  Of all the European languages, English is the most analytic, and it also contains many compounds.

      In synthetic languages words tend to be complex, consisting of content root morphemes with one or more affixes.  Most European languages are synthetic and use both prefixes and suffixes extensively, with English being the one that tends most toward the analytical bent.  Remember, the difference between synthetic and analytic is one of degree.  All languages use both techniques to some degree.

      Among synthetic languages there are languages like Quileute and Swahili which make exclusive use of prefixes and lack suffixes altogether.  There are also languages that lack prefixes but have numerous suffixes, such as one group of languages in northern Australia, as well as Mongolian and many other languages of Central Asia. 

      Languages where the words tend to be extremely complex in morphological structure are called polysynthetic languages.  In many polysynthetic languages a word may contain bound morphemes corresponding to both verb and noun in English.  This means that what are subject and predicate in an English sentence will often be expressed by a single word in a polysynthetic language. Here is a Nootka example:

inikw-ihl'-minih-'is-it-a (verb) inikw-ihl'-minih-'isit-i (noun)

fire-in house-plural-small-past ongoing
several small fires were buring in the house

fire-in house-plural-small-past ongoing-
the several small fires burning in the house

Since polysynthetic languages often link both subject and predicate into a single word, a process known as incorporation, they are sometimes called incorporating languages.  Many Australian languages and Native American languages are incorporating or polysynthetic languages; notable among them are the Eskimo-Aleut family, Algonquin languages of the northeast, and the Salishan languages of the Puget Sound, and also Chuckchi-Kamchatkan languages of Siberia. 

The difference between a word and a sentence becomes at times obscure, as in the Eskimo sentence-word angya-ghlla-ng-yug-tuq. He would like a big boat. (boat-suffix meaning large size-acquire-suffix expressing desire-3rd sing. verb suffix).  Here is another example from Tiwi, an Australian language:

 ngi-rru-unthing-apu-kani, I kept on eating.  (I-past tense suffix-a while-eat-repeatedly).


Languages of the world use three basic strucural techniques for showing syntactic relations.  Most languages use all 3 to some degree:

1) word order-- The big spider frightened the girl.

2) function words--The girl was frightened by a big spider.

3) inflections--Magna aranea perterruit puellam.

English:          The bad dog bit the man.

Chinese:         Gou3 yao3 ren2 le

Latin:   Malus canis hominem mordet.

Russian:         Zlaya sobaka ukusila muzhchinu.

      Some languages have many inflections to denote syntactic functions such as subject of the verb, direct object, indirect object, object of a preposition.  Finnish nouns take a total of fifteen separate endings to mark such syntactic functions.  These syntactic markers are called cases:  English only has 1 1/2 cases (genitive for nouns and pronouns, and object forms for pronouns only.) 

      There is a tendency that the fewer inflections a language has the more strict the word order.  English and Chinese very few inflections and also have strict word.  Latin has many inflections and free word order: Magna puellam perterruit aranea. Word order is then free to express stylistic nuances.  Russian is the same: Masha pokupaet knigu.  But like every other universal statement about specific structures in language, there are exceptions to this tendency, as well.  Lisu, a language spoken in Burma, is like Chinese and English in that it has no inflections to mark the relations subject and object, yet the language has free word order like Latin and Russian.  The question arises as to how this language expresses its syntactic relations without ambiguity.  Linguists are looking into it.     

Word order typology

      Apart from cases of free word order such as Latin, Navaho, or Russian, languages can be classified typologically according to the order each favors for expressing the syntactic elements subject, verb, object.  It turns out that over 75% of the world's languages put the subject first. 

The most common order is SVO (He killed the dragon) and is used by such disparate languages as Engl, Vietnamese, Hausa. 

The other, SOV (He the dragon killed), used by Japanese, Korean, Georgian, Cherokee, Tibetan and many other languages. 

Next comes VSO (Killed he the dragon) with 15%, including Welsh, Hawaiian, Squamish. 

VOS is even more rare less than10% (Killed the dragon he), and is used by Malagasy Tzotsil.

      But the rarest word order types are object initial languages.  In fact,  for many years object initial languages were not thought to exist at all; many linguists asserted that object-first languages violated a language universal requiring the subject or verb to go first.  But now several object first languages have been found in the Amazon basin. 

OVS (A dragon killed he.) is used by Hixkaryana and a few other Amazonian languages. 

OSV (A dragon he killed.) is used by Apurina and a few other Amazonian languages. 

      The classification just given is not absolute. The same language can have one dominant proclivity for a certain word order, yet also have sentences with different word orders. Remember that even in a predominantly SVO language like English, sentences with other word orders occur on a minor basis, usually as stylistic variants: 

SOV: pensive poets  painful vigils keep (Pope);

VSO: govern though my song (Milton);

OSV: What fools these mortals be (Shakespeare).

      The near total lack of languages favoring object initial word order may simply be a historical coincidence and say nothing about the human brain, which is creative and can devise an object-first language any time it wants.  It is interesting to note in this connection that in the movie Star Wars the character Yoda spoke English with the order OSV:  A sign you shall seeYour father he is.