Braine's 1963 analysis of two-work utterances in the light of a purely grammatical terminology led many critics at the time to belief that pivot grammar represented a fundamental, primitive rule regarding how children organise language. The simplicity of this formalism, however, eventually proved to be largely false, mainly because firstly, pivot grammar does not include semantic rules of acquisition, secondly, the rules are too rigidly applied and thirdly, attempts to expand the rules of pivot grammar lead to extremely inclusive conclusions that say very little about the specific manner in which two-word utterances occur in children. Braine's analysis was significant however, because it suggested that the syntactical rules of primitive acquisition of grammar in children did not follow the same rules as adults. The development of language and vocabulary in children, therefore, was seen to be fundamentally based upon the use of a pivot system whereby the repetition of pivot words could be used to expand the language of a child. Although the initial assertion of this rule's ubiquity was later reduced to a mere occasional trend in children, the assertion of a separate set of rules for children was influential in attempts to theorize a particular set of primitive grammatical rules. By contrast, Fillmore's 1968 theory of case grammar attempts to broaden Chomsky's rules of transformational grammar, as well as Braine's pivot grammar, by including the concept of case, and therefore, by looking at semantic acquisition of language, which Fillmore suggests sheds more light on the particulars of how language forms and develops in children. The results were much more complex and variable, and suggested that the primitive rules that govern language are heavily indebted to the context in which they occur. Although Fillmore continually asserted the existence of a series of primitive rules for language acquisition, the terminology of case grammar continues to prove difficult and ambiguous in many contexts. This essay will firstly discuss each of these theories, and then offer a comparison between these two theories.
Braine's (1963) analysis of language in infants suggests that the two-word utterances spoken by a child are not entirely based upon the random juxtaposition of words. Instead, Braine suggests that infants frequently distinguish between two separate word classes - the pivot (P) and the open (X) class of words. The pivot class consists of words that occur frequently and in fixed positions (at the end or the beginning of the enunciation). While these pivot words tend to vary between different children, they are often pronouns, such as "it", "that" or "my", prepositions such as "off" or "up", and certain verbs and adjectives, such as "do", "pretty" and "see". By contrast, Braine identified the "X-class" of words, later described as "open class", which occur in different positions and also occur with much greater frequency in one-word utterances. He hypothesised that early two-word combinations were comprised of either a type one pivot (which occurs at the beginning of an utterance) followed by an open class word, or an open class word followed by a type two pivot (which occurs at the end of an utterance). Braine posits in this paper that this rule could be one of the fundamental internal rules of grammar. Indeed, evidence of the ubiquity of this occurrence was corroborated by a number of other theorists at the time (Brown and Fraser 1964; Miller and Ervin 1964), who looked at instances in different circumstances where this grammatical rule would arise. However, eventually the theory was proven to have less import and its fundamentality was eventually demonstrated as inconsistent and far from ubiquitous across a wide range of children.
Braine's hypothesis has been widely criticised, both in terms of the terminological use of various components, and in terms of its more general approach and prioritization of syntactical form over semantic content in language acquisition. Firstly, Braine's theory has been refined, rarefied and modified by other linguists who see value in the general model of syntactical acquisition of rules but argue that the terminology used is too generalised. As Aitchison (1989) suggests: "Of course, there is nothing wrong with stating that some youngsters make sentences which can be P + O, O + P or O + O. It just does not tell us very much to say that, 'As well as pivot constructions, almost any other two words can occur together'" (115). Thus, terminologically Braine's theory is somewhat diffuse, as the principal rule is expanded upon its failure to encompass almost every possible outcome. Secondly, refinements to this theory, such as that of Brown (1973), have been criticised on grounds that these syntactical theories are far from universal across cultures (see Bowerman 1973), and when they are, they are so universal that they tell us very little indeed. In addition, the theory was also seen to neglect the importance of semantic content, preferring to rely instead upon a series of broad-ranging syntactic rules that neglected the importance of semantic meaning and context. Aitchison suggests that "it is not necessarily correct to assume that O + O utterances are random juxtapositions. There may be more reason behind them than appears at first sight, and the words may be related to one another in a highly structured way." Bloom (1970) suggests that the focus on the syntactical elements of speech in this instance "fails to capture the semantic richness of these simple utterances" (Harley 2001, 122). Thus, while Braine's theory was significant insofar as it highlighted different grammatical rules in operation in children's acquisition of language and grammar, it is limited by both the terminology used, which is highly generalised, and by the lack of a formalised way of measuring the semantic or context based acquisition of language.
Fillmore's case grammar (1968) was developed as a response to Chomsky's model of transformational grammar which, he argued, did not cover the functional aspect of clause items as well as their category. As an example, Chomsky's transformational grammar did not identify a difference between "the child opens the door" and "the key opens the door", as both statements are classified as subject-verb statements. By contrast, Fillmore's case grammar looks at the case, or the semantic function of these terms, and asserts a fundamental difference based upon these principles; in this case, the former would be an example of an agent-action-object sentence, while the latter could be classified as an instrument-action-object sentence. Fillmore (1968) identifies the categories as follows: agentive (A); instrumental (I); dative (D); factitive (F); locative (L); and objective (O) (24-5), with the possible addition of benefactive and comitative cases (81). Fillmore argues that these case grammars are primitive as opposed to derived - as in, they occur and develop in certain ways in children when they make two-word utterances. An important extension to Fillmore's work is Bowerman (1971), who argued against pivot and Standard Theory grammars because they fail to look into the context of the statements that are used. Bowerman argues that grammatical case is derived from the mother of the child concerned; thus, while Standard or pivot-based grammatical systems would see these utterances as nonsensical, and / or derived from a fundamental rule of grammar based upon grammatical acquisition alone, the inclusion of case suggests that grammar in these early cases are not nonsensical and do in fact follow patterns of case derived from the mother. As Ingram (1989) suggests, Bowerman's study of Seppo, a Finnish child, demonstrates that "the mother's word order is probably the source of the consistent order used by Seppo" (282). Of course, the development of a case-based grammatical system leads to its own terminological problems; it is, however, more useful than case grammar because it points to the semantic acquisition of language, and does not merely treat the content of two-word utterances as nonsensical when they do not abide by certain grammatical rules. The difficulty with case grammar is that the terminology described above, coupled with a context-based and therefore relatively subjective model for the acquisition of language makes the appropriation of a universal series of grammatical rules more difficult to come by. On the one hand, this can lead to a criticism of the lack of taxonomic rigour in the system, but on the other hand suggests a greater versatility and adaptability in describing the development of certain rules in the development of language in children. Indeed, Bowerman (1973) suggests that the simplicity of pivot grammars "fail by assigning more structure to the child than Bowerman feels is warranted by the data" (Ingram 281). The initial theoretical assertion of the ubiquity of pivot rules in early infant development of language eventually proved to be a significant shortcoming in pivot grammars, as these rules were later seen as being far from ubiquitous - thus the validity of the theory and its structural basis has little validity when applied to the actual development of two-word utterances of children, which were seen by Fillmore and Bowerman as deeply contextual and built upon the acquisition of linguistic traits from social, rather than primitive, origins. Indeed, the application of universal rules in the acquisition of grammar in children proves to be a serious shortcoming in most of these formalisms, although Fillmore's inclusion of context allows greater scope for linguistic analysis to apply to less rigid and simplistic patterns of primitive acquisition.
Both Braine (1963) and Fillmore (1969) assert the primacy of linguistics over psychology by asserting the presence of basic linguistic rules that govern the development of language. However, the ever-changing nature of these rules, formalisms and taxonomies, along with the continual disclosure of the falsity of these rules have made formalism more difficult to ascertain in linguistics. While both Braine's pivot analysis and Fillmore's case grammar have their strengths and weaknesses, the latter can be said to have usurped the former because there is greater room within its particular rubric for interpretation and the incorporation of what we increasingly suspect to be one of the only rules to which we can genuinely rely upon: that the convergence of primitive and social acquisition of grammar can never be wholly correct. The central difficulty of Braine's analysis was precisely that pivot grammar was constructed as a theory that was universally applicable to all children in two-word utterances. Thus, the theory was too rigid and simplistic to incorporate changes easily. Fillmore's case grammar, on the other hand, allowed a greater scope for analysis; the greater complexity of taxonomy, as well as allowing a greater degree of versatility, insofar as it was more inclusive of semantic, as well as syntactic rules of acquisition, also allows for a greater level of insight into the psychological side of psycholinguistic analysis. Thus, while Braine's pivot grammar allows us to look into the syntactic elements of two-word utterances, Fillmore's case grammar is a more durable taxonomic system from which greater insights can be gleaned.