Chapter I. Lexical typology is the systematic study of cross-linguistic variations

1.3. Systematic study of cross-linguistic variation
The distinction between count and mass is the main emphasis of this course's discussion of cross-linguistic variance in semantics. The way nouns like pen and water mix with, say, numbers have implications for the difference between countable objects (like pens) and non-countable things (like water); consider two pens vs. two cups (or liters, etc.) of water. There is a great deal of inter-linguistic variance when examining how languages encode countability. For instance, English count nouns ending in -s, such as pen, combine with a numeral directly (two pens). In Mandarin, a "numeral classifier" (CLF) identifying the unit of counting must be added before the word "pen" can be used with a numeral. Mandarin's translation of the English phrase "two pens" is "two CLFbranch pens," which literally translates to "two branch pens." The term "CLFbranch" denotes the fact that long, thin objects are being counted. The Mandarin term for "pen" is similar to English mass nouns like "water" in this regard: when employing a numerical with a mass noun, an expression like as "cup" that denotes the unit of counting must be added, as in "two cups of water."
Other languages merely say "two pen" instead of using a classifier or plural indication (Tagalog, Yudja). Or they permit other structure, such as "two pen," "two unit pen," or "two pens" (Indonesian, Armenian). This looks at semantic ways that cope with this variance on the one hand and language variation in the count/mass distinction on the other. We will also explore the issue of how the patterns we discover relate to human mind.
Students can opt to concentrate on language description and typology or formal semantic features. Different points of departure underlie claims about universals in language acquisition. A nativist, Chomskyan-based view will specify these in terms of the formal principles characterizing a generative model of Universal Grammar (UG) with cross-linguistic variation accounted for by differential setting of structural parameters according to the target language. A Piagetian orientation will explain them in terms of general cognitive principles and shared conceptual underpinnings which provide a universal epistemological basis for acquisition of linguistic knowledge (Slobin 1985). The view espoused by this author is that children start out the task of acquisition equipped with categorial and semantic distinctions of the type termed by Chomsky ‘substantive’ as against formal universals, akin to Keenan's (1979) idea of ‘naive’ or primitive universals (Berman 1993b). For example, children do not need to learn that there are two main kinds of sound segments, consonants and vowels; that propositions distinguish between predicates and arguments; or that there are intransitive and transitive clauses, with one-, two-, or three-place predicates.
Whatever the explanation they propose, researchers agree that there is a clear biological underpinning to the overall developmental stages manifested by children in command of language production irrespective of mother tongue. Just as children sit before they stand, and walk before they run, so children across the world babble before they produce single words, rely on single words before they start combining words, and produce isolated clauses before they join these together by increasingly specific lexical means and complex syntactic devices. Overall progression thus involves an evolution in construction types. Language production typically starts out with utterances, as behavioral elements which provide the initial basis for expressing relations between semantic content and phonetic form. These typically take the form of isolated lexical items or unanalyzed strings of words, starting around the end of the first year of life. Structure dependence emerges when these lexical units are combined in simple-clause structure, realized by morphosyntactic markings of verb-argument relations. Next, adjacent clauses are combined, first by coordination, subsequently by various kinds of subordinate constructions—initially mainly finite and subsequently nonfinite—leading up to complex clause-linkage by derived nominalizations (Berman 1993a). This needs to be qualified for a verb-final language like Turkish, where the unmarked means of clause linkage is by nonfinite, nominalized constructions. These emerge relatively early in Turkish compared with European or Semitic languages.
With respect to acquisition of morphology, the domain in which languages tend to differ most, children generally acquire the bulk of the inflectional system of their language before they master its derivational morphology. This is understandable since inflection is more regular, productive, and grammatically obligatory than derivation. Besides, derivational processes depend on a level of vocabulary expansion not achieved by young preschool age. In fact, in some instances, like derived nominals or the high-register. Latinate lexicon of English, children will master derivational processes only well into school age. In contrast, they by and large command the inflectional systems of their language by the age of three, irrespective of how morphologically opaque or synthetic these happen to be. Cross-linguistic differences affect acquisition of inflection along two main dimensions. Psychological processing principles like perceptual saliency coupled with regularity and transparency of form-meaning mappings account for the fact that, for example, Turkish-speaking children gain command of the bulk of their rich system of agglutinating, inflectional suffixes by a much earlier age than children acquiring languages with more synthetic or syncretic morphologies like Hebrew or Icelandic. And experiential factors of frequency of use in the everyday colloquial speech that constitutes the linguistic input and output of young children explain why certain grammatical categories are acquired relatively late in languages where they constitute high-register or literate constructions.
In the domain of syntax, children typically gain command of simple clause structure in their language between ages two to three years. Basic word order emerges very early indeed, first between phrasal constituents, only later within phrases. Constituent ordering is followed by grammatical marking of verb-argument relations, around the same time or somewhat before inflectional marking of grammatical agreement. The overall progression is thus much as Brown (1973) defined for English: from single words to early word combinations, thence to simple clause structure, followed by a range of devices for clause linkage.
Another commonality is the general development observed across children from rote-learning of unanalyzed items to partial productivity followed by command of structure-dependent rules and generalizations, which are subsequently constrained by conventions of lexical and discourse usage. This has been demonstrated, for example, for how children acquire the semantics, morphology, and syntax of causative verbs in three typologically distinct languages: English at the more analytic or syntactic end of the scale, Hebrew with relatively synthetic bound morphology, and Inuktitut as a polysynthetic language which marks causatively by both morphological and lexical means (Allen, in Lieven 1998).General language development thus proceeds across different phases, from programmatical to structure-dependent to appropriate use. And it becomes both more elaborated and varied in structure and use, on the one hand, and more restricted or constrained by target-language specifics, on the other. Thus, in acquiring the sound system, babies proceed from universal crying to uttering a range of speech sounds, followed by babbling, which is increasingly restricted to elements of the native language to a point where eventually the system will be so constrained as to impose a non-native pronunciation in other languages. Similarly, in morphology, across languages children proceed from no alternation between forms to immature strategies for initial markings of grammatical categories, thence to ‘creative errors’ of over-regularization, until eventually the system becomes fully constrained by lexical convention. And in the acquisition of the lexicon, again across languages, children move from a restricted vocabulary of nursery terms and over or under extensions of conventional items to a period of creative innovation; eventually the conventional lexicon takes over, while at the same time vocabulary growth is both constrained and elaborated by socially and culturally appropriate distinctions of register and usage, lasting into adolescence.
Both parameters of cross-linguistic variation discussed so far are related to inflection. The question arises whether all parameters can be stated in terms of inflectional variation. One area of cross-linguistic variation whose definition remains subject to controversy concerns the contrast between so-called VO languages such as English, in which the verb precedes its direct object, and OV languages such as Dutch and German, in which the direct object precedes the verb.