Desktop version

Home arrow Language & Literature


Within a classical WP model, words not only function as basic lexical units but also play a pivotal role in what is often termed ‘the creative use of language. The description of principal parts in Hale and Buck (1903), summarized in Section 4.3.2, sets out how the information expressed by exemplary paradigms and principal parts is combined to define novel forms.1 The special status of an exemplary paradigm lies in the fact that it exhibits general patterns of inflection. A set of principal parts contributes item-specific word forms. Matching the principal parts of an item against cells of an exemplary paradigm establishes a correspondence between principal parts and their counterparts in the exemplary paradigm. New forms are then deduced by generalizing this correspondence analogically to cells whose forms have not been encountered.

In contemporary grammatical models, the relevance of analogy is usually taken to be confined to pedagogical and historical contexts. To the extent that analogy is invoked in synchronic processes, its application is restricted to ‘local’ generalizations that run counter to more general rule-governed patterns. However, from a classical WP perspective, analogy guides language acquisition, use and development. Although analogy figures in one of the earliest recorded linguistic debates, the analogy vs anomaly controversy of classical antiquity, it is in the Neogrammarian period that analogy is most explicitly associated with creative and productive language use. The role of analogy in the acquisition and use of morphology is forcefully expressed by Paul (1920):

The creative activity of the individual is also very considerable in the domain of word building and even more so in inflection ... We see the effect of analogy especially clearly in the grammatical acquisition of inflected forms of a foreign language. One learns a number of paradigms by heart and then memorizes only as many forms of individual words as is necessary to recognize their affiliation to this or that paradigm. Now and then a single form suffices. One forms the remaining forms at the moment that one needs them, in accordance with the paradigm, that is, by analogy.2 (Paul 1920:112)

  • 1 Here and below, ‘novel forms’ refers to unencountered forms, not neologisms.
  • 2 Sehr bedeutend ist die schopferische Tatigkeit des Individuums aber auch auf dem Gebiete der Wortbildung und noch mehr auf dem der Flexion ... Besonders klar sehen wir die Wirkungen der Analogie bei der grammatischen Aneignung der Flexionsformen einer fremden Sprache. Man lernt eine Anzahl von Paradigmen auswendig und pragt sich dann von den einzelnen Wortern nur soviel Formen ein, als erforderlich sind, um die Zugehorigkeit zu diesem oder jenem Paradigma zu erkennen. Mitunter genugt dazu eine einzige. Die ubrigen Formen bildet man in dem Augenblicke, wo man ihrer bedarf, nach dem Paradigma, d.h. nach Analogie.

Word and Paradigm Morphology. First edition. James P. Blevins © James P. Blevins 2016. First published 2016 by Oxford University Press

A similar outlook is expressed by Hockett (1967:221) when he imputes “a net gain in realism” to “the student of the language” who “would now be required to produce new forms in exactly the way the native user of the language produces or recognizes them—by analogy”. Matthews (1991) likewise illustrates the process of analogical generalization with reference to the inflection of the Latin nouns dominus ‘master’ and servus ‘slave’:

In effect, we are predicting the inflections of servus by analogy with those of dominus. As Genitive Singular domini is to Nominative Singular dominus, so x (unknown) must be to Nominative Singular servus. What then is x? Answer: it must be servi. In notation, dominus : domini = servus : servi. (Matthews 1991:192Q

In the continuation of this passage, he echoes Paul’s position in attributing the celebrated ‘U-shaped learning curve’ to the effects of analogy:

Analogy is an important concept in linguistic theory. It plays a major role in morphological change, as we have noted earlier. It also forms a large part of the process by which children learn their native language. One of the most banal and often repeated observations of children’s speech concerns the extension of regular inflectional patterns (English -ed, -s, and so on) as analogical replacements... It is hardly surprising that traditional language teaching has made use of the same instinct. (Matthews 1991:192f.)

Processes of analogical extension are implicit in earlier discussions of the use of partitive singulars to deduce novel forms of Estonian nouns or the use of nominative singulars to determine the novel forms of Russian nouns. Similar principles underlie the use of conjugation tables to exhibit patterns of verb inflection. Verbs in Russian, Estonian and Latin have larger paradigms and larger principal part inventories, but the processes remain the same.

The deduction of forms from exemplary patterns is standardly formalized in terms of what are known as proportional analogies. The canonical format is a ‘four-part analogy’ of the form a : b = c : X, in which a, b and c are all given, and the analogical step involves ‘solving for X’.[1] Matthews’ deduction of servi in the passage directly above illustrates the form and application of four- part analogies. As this passage clarifies, the apparent simplicity of the traditional formulation rests on tacit assumptions about the terms it contains. The antecedent terms a and b do not represent simple forms, but instantiated cells in an exemplary paradigm, ([nom,sg],dominus) : ([gen,sg], domini) in Matthews’ example. The relation between these instantiated cells provides a model for generalizing from a principal part c, here ([nom,sg], servus), to the form X in the corresponding instantiated cell ([gen,sg], X) in the paradigm of servus. These interpretations are unpacked in (5.1), in which F and G represent feature bundles and ф, фг, ф and ф' represent forms.[2]


  • 99
  • (5.1) Interpretation of terms in traditional four-part analogy a : b = c : X

ф) : ф') = ^) :

Four-part analogies represent the smallest proportional deduction, as at least three known forms are needed to identify a fourth. The exemplary forms a and b exhibit a pattern, which is extended from the principal part c to the unknown form X. Without both a and b, there is no pattern to extend, and without c no base for the extension. A deduction may also specify more than two exemplary forms when any single form matches multiple patterns. Paul (1920:107) gives triples such as Tag: Tages : Tage = Arm : Armes : Arme to illustrate proportions between case forms of the German nouns ‘day’ and ‘arm’.

As a member of the last generation of linguists to be trained in the Neogrammarian tradition, Bloomfield retained a notion of regular analogy, which he also accorded a central role in the creative use of language:

A grammatical pattern (sentence-type, construction or substitution) is often called an analogy. A regular analogy permits a speaker to utter speech-forms which he has not heard; we say that he utters them on the analogy of similar forms which he has heard. (Bloomfield 1933:275)

It is only during the Post-Bloomfieldian period that analogy came to be reinterpreted as a minor process governing idiosyncratic deviations from productive patterns based on superficial similarities. Chomsky (1975) contains a blanket dismissal of familiar analogical processes, including “inductive procedures, methods of abstraction, analogy and analogical synthesis, generalization and the like” on the grounds that “the fundamental inadequacy of these suggestions is obscured only by their unclarity” (Chomsky 1975:31). Chomsky’s assessment of analogy, and the motives of those who invoked this notion, become more scathing over time, as objections to “inadequacy” and “unclarity” gave way to charges of outright “misleading” equivocation:

Suppose one were to argue that the knowledge of possible words is derived “by analogy”. The explanation is empty until an account is given of this notion. If we attempt to develop a concept of “analogy” that will account for these facts, we will discover that we are building into this notion the rules and principles of sound structure. There is no general notion of “analogy” that applies to these and other cases. Rather, the term is being used, in an extremely misleading way, to refer to the properties of particular subsystems of our knowledge, entirely different properties in different cases. (Chomsky 1988:26f.)

The idea that analogy amounts at best to an imprecise proxy for generative rules is expressed more directly by Kiparsky (1975:75) when he claims that “generative phonology argues that at the point where the analogies begin to make the right generalizations, they are indistinguishable from rules”. Following a discussion of processes of analogical levelling and extension, Haspelmath (2002:56) likewise concludes that “the solution of an analogical equation is practically the same as the application of a word-based rule to a novel word”. Generalizing this point to express a consensus that spans the Chomskyan tradition and extends to some degree beyond it, Haspelmath (2002:103) goes on to suggest that “some morphologists have concluded... that morphological analogy and morphological rules are really one and the same thing”.

  • [1] Although proportional analogies are often associated with Paul (1920), Morpurgo Davies(1978:46) notes that they appear slightly earlier in the work of Paul’s contemporaries.
  • [2] The antecedents a and b are usually drawn from the same exemplary paradigm, but this reflects ause of four-part analogies rather than a property of the format.
< Prev   CONTENTS   Source   Next >

Related topics