You are on page 1of 5

Chapter 2

Typology and universals

Typology is the study of linguistic systems and recurring patterns of linguistic sys-
tems. Universals are typological generalizations based on these recurring patterns. This
chapter will give a definition of what typology is (2.1). Section 2.2 brings up the notion
of classifications, especially genealogical, and gives a brief discussion on why pidgins,
creoles and mixed languages, as well as sign languages tend to be treated separately
(2.2.1 and 2.2.2 respectively). In 2.3 I give a definition of universals and bring up the
concepts of unrestricted (2.3.1) and implicational (2.3.2) universals. Section 2.4 takes a
cursory glance at the motivations for language universals.

2.1 What is typology?

To put it very simply, linguistic typology concerns itself with the study of structural
differences and similarities between languages. The term typology is, as many other
linguistic terms, borrowed from the field of biology and means something like ‘tax-
onomy’ or ‘classification’ (Croft 2003: 1), or, to be precise, “the study and interpreta-
tion of types” (Pearsall & Trumble 1996: sv). Linguistic typology, then, is the study
and interpretation of linguistic or language types. More specifically, it is the study
and interpretation of types of linguistic systems. While this may involve comparison
of linguistic systems within a language, it more generally involves comparison of lin-
guistic systems between languages. Linguistic typology can be both synchronic, i.e. a
comparison of languages contemporary to each other, or diachronic, i.e. a compari-
son of languages at various stages of their historical development. Impressionistically,
synchronic typology has received more attention than diachronic typology, but both
are equally necessary and can be thought of as complementary to each other (cf. the
discussion in Croft 2003: 232ff).
Any linguistic system may serve as a starting point for typological comparisons.
Thus we may, for example, have typological surveys of phonological, morphological,
grammatical, syntactic, lexical, pragmatic, semantic, etc. systems. Essentially, the ulti-
mate goal of linguistic typology “is to understand [the question] ‘what’s where why?’ ”
(Bickel 2007: 248). In other words, a driving force is to try to establish recurring pat-
terns across languages, in order to answer the questions “what is out there?”, “where
does it occur?” and “why do we have particular patterns?”. If we want to formulate
16 Introduction to Linguistic Typology

hypotheses about the unity, diversity, potentials and limits of human language, we
need to know what human language is capable of. Investigating only one language
will not be sufficient to answer such questions. For instance, if we look at English we
may establish various linguistic factors, such as what the phoneme inventory is, what
the morphology is like, what grammatical categories we can discern, how units are
ordered, and so on. We may then use these factors to hypothesize what the human
brain needs or does not need in order to allow a person to produce and maintain a
coherent language and communicate with others. Based on English, this might lead
us to assume, for example, that a human language needs quite a number of vowels in
order to get by. We might further assume that the only way to know what grammatical
functions the words in a sentence have depend on how they are ordered with respect
to each other. Compare, for example, the following:
(4) a. John called Mary.
b. Mary called John.

In (4a) John is the person who did something (John is the subject of the clause), and
we know that because John is placed before the verb called. Mary is the object of the
clause because Mary is placed after the verb. If we swap John and Mary the grammati-
cal relations also swap and Mary becomes the subject while John is the object, as in
(4b). Based on English, we would thus assume that word order follows a rigid pattern
of subject-verb-object. We might also assume that a subject must be expressed in a
clause for the clause to be grammatically acceptable, even if there is no physical entity
to be referred to. In a sentence like He swam, for example, He refers to some (male)
human or animal. He is the subject of the verb swam. But in a sentence like It rained,
we have a neutral pronoun filling the slot of the subject, even though It does not refer
to anything. Taking out the subject, however, is not possible; a clause like *Rained is
not grammatically acceptable.4 This might lead us to conclude that the human brain
demands that every clause has at least one slot for “subject” and one for “verb” in order
to be complete, even if the “subject” slot is filled with a semantically empty reference
(i.e. even if the reference does not have any concrete real world meaning).
Based on some other language, we might get an entirely different picture, which
would lead us to make an entirely different set of assumptions. If we base our hypoth-
esis on what a human language must have on Mandarin, for example, we would assume
that we only need a handful of vowels to get by, but that differences in tone is an
absolutely essential requirement of human language. Compare the four words in (5):

4. The symbol * means that the form or construction is non-existent or grammatically unacceptable.
Chapter 2. Typology and universals 17

Mandarin (Sino-Tibetan (Chinese): China)


(5) a. yī ‘cloth’ b. yí ‘to suspect’
c. yiˇ ‘chair’ d. yì ‘meaning’ (Li & Thompson 1990: 814)

As (5) shows, the choice of tone (indicated on the vowel by the diacritics ˉ, ́, ˇ and ̀
respectively) is an essential part of conveying the meaning of the word in Mandarin.
Furthermore, if we base our hypothesis on Mandarin only, we might assume that copy-
ing parts of the word, a process called reduplication, is an essential requirement for a
language to be able to form adverbs from adjectives:
Mandarin (Sino-Tibetan (Chinese): China)
(6) màn ‘slow’ → màn-màn-de ‘slowly’ (Li & Thompson 1990: 821)

Both tone and reduplication are largely irrelevant for the linguistic system of English.
Based on yet another language we would make yet different assumptions. In
Modern Greek, for example, the order of the elements in the clause is not rigid.
Compare the sentences in (7) below:
Modern Greek (Indo-European (Greek): Greece)
(7) a. o skilos kiniyinuse ti yata
the dog chased the cat
b. kiniyinuse o skilos ti yata
chased the dog the cat
c. o skilos ti yata kiniyinuse
the dog the cat chased
d. ti yata o skilos kiniyinuse
the cat the dog chased
e. ti yata kiniyinuse o skilos
the cat chased the dog
f. kiniyinuse ti yata o skilos
chased the cat the dog
‘The dog chased the cat.’ (Ruge 1984: 94)

In (7) the elements of the sentence (the dog, the cat and chased respectively) may move
around in relation to each other. The role of subject always stays with o skilos ‘the
dog’, because o skilos is morphologically marked for the nominative case. The role of
object always stays with ti yata ‘the cat’ because ti yata is marked for the accusative
case. Based on Greek we would therefore assume that while the order of the elements
in the sentence is largely irrelevant, morphological devices to indicate the case of the
nouns are essential tools for humans to be able to sort out what grammatical relations
18 Introduction to Linguistic Typology

the elements have. Case marking is largely irrelevant for English and Mandarin.
Furthermore, based on Greek, we would conclude that an overt reference to a subject
is not necessary if the subject is not a real world entity:
Modern Greek (Indo-European (Greek): Greece)

(8) θa vréksi
fut rain.3sg
‘(It) will rain.’ (Joseph & Philippaki-Warburton 1987: 23)

In (8) there is no noun or pronoun referring to ‘rain’ or ‘it’ (θa is a marker for tense
and belongs to the verb). In Greek, only an inflected form of the verb ‘rain’ is neces-
sary for the sentence to be grammatically correct. This would not be possible in either
English or Mandarin.
What I have tried to show here is that if we look only at one single language, or
maybe a very small number of languages, and base our hypotheses on what humans
require, or tend to need, or tend to avoid in their communicative tool called language,
we are likely to end up with assumptions that would not hold, given that languages can
differ from each other a great deal. Looking at a larger number of languages we would,
for example, see that, contrary to English, languages can easily get by with very few
vowels, or that some languages demand a very large inventory of consonants indeed to
differentiate between different meanings. We would find that some languages employ
very little or no inflectional morphology, while others demand extremely complex
kinds of inflections for sentences to be grammatical. We would see that while some
languages have a rigid word order, others do not. In other words, if we want to be able
to answer the fundamental question “what is language?” then we will have to know
what kinds of solutions speakers have found to encode meaning, i.e. what kinds of
systems their languages have ended up acquiring. In order to get hold of such informa-
tion, we need to compare between languages across both space and time.
Linguistic typology thus often involves cross-linguistic comparison, i.e. com-
parison between different languages.5 While the term ‘cross-linguistic’ as such simply
means “across languages” and can be used for a comparison between only two lan-
guages, I use it here and throughout this book to imply across several languages. With
cross-linguistic comparison linguistic typology can give an idea of how linguistic fea-
tures pattern across the world. Some solutions are common, some are rare. Typology

5. This is often contrasted with Generative Grammar, which approaches the study of language by
looking very closely indeed at one language or a small number of languages. Different approaches to
the study of language all have their merits and complement each other in the common quest to try
to understand the fundamentals of human language. For an accessible introduction to Generative
Grammar, see, for example, Carnie (2002).
Chapter 2. Typology and universals 19

can map the patterns, which may then serve as a starting point for investigations into
why we find those kinds of patterns. Typology can also serve as a guide to analysis of
languages.
Once there is a clear and precise classification of occurring patterns, a new pattern
may be evaluated with respect to existing ones. In diachronic [i.e. historical] analysis,
where pieces of the puzzle (living speakers, phonetic studies of them, etc) may often
be missing, typological work can be particularly useful in guiding the analysis in one
direction over another. (Blevins 2007: 110)

While the quote above focuses specifically on the importance of typology for explain-
ing sound patterns, the statement holds for any linguistic domain.
It is important to note here that typology is not able to establish what is possible
in human language, for the simple reason that not all human languages are available
for analysis. For one, it is not possible to include all of the roughly 7000 languages
currently known to exist in the world (Lewis 2009), because we do not have adequate
descriptions for all of them. But even if we did have descriptions for all the known
languages in the world, we do not have access to all languages that have already gone
extinct, nor do we have access to those languages that are yet to emerge. Thus typology
is about probabilities more than anything else. In short, linguistic typology is basically
about tendencies and by extension trying to explain why we get these tendencies.

2.2 Classifications

A central concept in typology is classification. Languages can be classified according


to various criteria. We could choose to group languages by the number of speakers they
have, or by formality of the situation where the languages are used, or by the area they
are spoken in, or by the genetic affiliation they have, and so on. With linguistic typol-
ogy the classification is primarily based on the elements that make up the structure of
languages, such as sounds, words and parts of words, how these words are organized
with respect to each other, and so on. Thus we could say that Mandarin is a reduplicat-
ing language (while English is not) and is also a tone language (while English is not).
We can say that both English and Chinese are subject-verb-object languages. And so
on. When investigating correlations between languages, we may, for example, investi-
gate whether reduplicating languages also generally tend to be tone languages (they do
not, in fact; most of the reduplicating languages in Rubino’s (2011) database are listed
as languages lacking tone in Maddieson’s (2011f) database). Since the bulk of this book
is about what types of linguistic systems and patterns we tend to find across languages,
that kind of classification will not be discussed further in this section.

You might also like