Book review

Analogy as Structure and Process

Itkonen, Esa. 2005. Analogy as Structure and Process. Approaches in linguistics, cognitive psychology and philosophy of science. Amsterdam: John Benjamins

Reviewed by Thomas Verjans (Université Paris 4-Sorbonne)

In this book, Esa Itkonen argues that "analogy plays a central role in human thinking" (XI). He distinguishes between analogy as structure and analogy as process, and he shows how the concept of analogy is used - or should be used - in linguistics as well as in other cognitive sciences.

Whereas the concept of analogy seems to have been neglected over the last few decades (see Marchello-Nizia 2006: 78), Esa Itkonen's new work aims "to show - in the spirit of Humboldt's 1812 and Whitney's 1875 dictum - that analogy is the central concept of language and linguistics" (6). This book contains a short introduction and five chapters (the last one being the conclusion), to which an annex is added. The first chapter offers a definition of the concept (chapter 1), while the subsequent chapters discuss a number of domains in which analogy can be applied (chapters 2, 3 and 4).

In the first chapter, "The concept of analogy", Iktonen defines the prototype of analogy as "structural~functional" analogy (2). Other kinds of analogy can be located on a continuum, the last step being formal analogy. He distinguishes two kinds of analogy: analogy as structure and analogy as process. He first illustrates the context-dependency of analogy, as well as its dependency on point of view. He then establishes "a taxonomy of the relations between two analogous systems" with respect to their "ontological and/or epistemic status", each of them symmetric or asymmetric. The four resulting types are exemplified by discovery, invention/creation, and imitation.

Then, he shows how the concept of analogy is related to other cognitive operations. He defines 'generalization-1' as an extension of a specific response and 'generalization-2' as the discovery of "invariant features common to A and B". Itkonen argues "that the concept of analogy is apt to clarify how the two senses of 'generalization' are related" (22). Next, he calls 'analogical extension' "the generalization-1 based on generalization-2" (23). He explains the relations between dynamic analogy and principal types of inference, particularly the abductive one, which is in a relative reversal relation with analogy: if the dynamic analogy is "a subtype of abduction", "every abduction is in turn more or less analogical" (35). The next section concerns the relations between analogy and metaphor on one hand, and analogy and blending on the other. The latter could be understood either as "just one exemplification of analogy among many other" (43) or as "a notion only contingently related to analogy" (44). Metaphor as well as blending are finally characterized by the following proposition: "all metaphors are analogies, but not all analogies are metaphors" (41).

After that, he discusses "analogy as a psychologically real and causally effective phenomenon" (50). Itkonen divides human behavior into instrumental actions, which are free and can be explained by a "rational explanation", and institutional actions, which are rule-governed. Each of them correspond respectively, in synchronic linguistic, to morphosyntax and semantics on the one hand, or pragmatics on the other. Itkonen defines analogical ambiguity as a competition between "several solutions in any given problem situation" (59).

In the last section of the chapter, he explores the necessary limits of analogy. Analogy is said to "rest on something non-analogical" (63), like innate concepts, and to be distinguished from other operations, like those based on association by similarity or contiguity.

The second chapter, "Analogy Inside Linguistics", explores how this concept can be used on an epistemological as well as on a methodological level. Itkonen rejects some objections against the concept of analogy that pertain to both synchronic and diachronic linguistics.

Itkonen highlights the role of analogy in different linguistic levels, first in phonology and morphology, in which "units, i.e. sounds and (inflected) words, are defined by their (analogical) relations to other units" (86), thus forming a coherent system. Moreover, he states that in morphology, "the use of analogy … should … be regarded as psychologically realistic" (85) as a process that facilitates memorization. Since the application of analogy to the syntactic level is more complicated, Itkonen exemplifies this aspect with three different traditions: the Arabic tradition, the generative tradition, and the paradigm of construction grammar. He shows that, in each of them, analogy plays a central role not only in the methodological description but also in the theoretical construction. Finally, on the semantic level, Itkonen defines iconicity as "vertical (directional) analogy, or analogy that goes from non-linguistic to linguistic", but he adds that "there is within semantics also 'horizontal' (directional) analogy" (103). The former corresponds to "ontological motivation" whereas the latter corresponds to "morphosyntactic motivation" (104).

Subsequently, Itkonen considers analogy in diachronic linguistics as well as in linguistic typology, e.g. in relation to sign languages. At first, he shows that, even if it is traditionally applied at a low-level, like morphology or syntax, "also such large-scale changes as affect the entire structure of a given language have to be subsumed under analogy" (108). After that, he discusses the role that analogy plays in grammaticalization, which is claimed as similar or analogous to the hypothetico-deductive method. Moreover, reanalysis, which is the first stage, can be highlighted by analogy in the sense that it explains its emergence toward a model. Thus, reanalysis is justified by the economical principle. Then Itkonen shows that "in spite of huge differences as to their respective physical bases, there is a fundamental analogy between oral languages and sign languages" (124). He concludes this chapter with an analogy-based view of language acquisition.

The purpose of the chapter 3, "Analogy and / or overlap between language and other cognitive domains" is to "investigate what linguistic analysis (as distinguished from psycholinguistic or neurolinguistic analysis) can reveal about the nature of the (human) mind" (129).

At first, Itkonen reconsiders iconicity to show that the modularity-hypothesis is refuted by it. Moreover, Itkonen defines his position vis-ŕ-vis the 'imagery debate' as the assumption for a single code, since he views language as "a 'picture' of" (136) reality.

Then, whereas the relation between language and vision is considered an analogical one, which is supported by the "ultimate unity of all types of perceptual experience" (137), and moreover if sign language is included as a subtype of language, the relation between language and music and language and logic is to be characterized by their common origin. In the first case, language (including sign language) and music are said to be derived "from a common source" (148) synchronically as well as diachronically. In the second case as well, "the common origin of language and logic" (162) is shown, the latter being created with the former.

He concludes in Jackendoff's (1994: 60) formulation, "language shares the general character of other faculties of the mind" (163). But, in contrast to Jackendoff, and moreover with regard to the analogical relations which are revealed by the preceding explorations, it signifies that it is impossible to infer from it that the mind is structured in a modular fashion.

In chapter 4, "Analogy (mainly) outside linguistics", Itkonen discusses the use of analogy in areas like mythology and cosmology as well as in scientific discovery. He argues that "the line of demarcation between these two types of thinking is much less clear-cut than one might perhaps expect" (165).

In the first case, the "binary classification" which is characteristic of the 'primitive' world-view is based on analogy, because structural similarities are related through analogy. However, this kind of "classification remains horizontal in the sense that all elements are situated at the same degree of abstraction" (173). But, "once the vertical dimension (…) has been brought in, we have the beginnings of a scientific way of thinking" (id.).

Discussing scientific discourse, Itkonen cautions against a restriction of the role of analogy only at the discovery-stage. He argues that analogy is within the generalization-stage as process (dynamic analogy) as well as result (static analogy).

He finally affirms (Chapter 5: "Concluding Remarks"), in agreement with von Humboldt and Withney, that "analogy is 'the lifeblood of human thinking'" (199). Subsequently, the Appendix, writen with Jussi Haukioja, is a continuation of the second chapter, and more precisely of the syntactic application of analogy. It presents the computer program PROLOG and offers a technical explanation.

Evaluation

Esa Itkonen's book is not only an in-depth account of the concept of analogy, but also an epistemological treatise that confronts scientific (linguistic or other) theories. In this sense, the distinction between the two kinds of analogy constitutes a very helpful evaluation criterion.

Finally, a few remarks should be added. First, two studies that also illustrate the importance of analogy should be seen as complementary to this work. P. Monneret's work (2003, 2004), which demonstrates how the analogy constitutes not only a central concept in linguistics and other domains, notably in a kind of linguistics that he calls "linguistique analogique". His Essai de linguistique analogique complements Itkonen's work. Another reference that is relevant to Itkonen's claims is the diachronic application of analogy explained in C. Marchello-Nizia's six-way typology (2006: 93-95). This work touches on phenomena like popular etymology or other subtypes of analogy as different kinds of linguistic change.

Finally, a potential criticism of the work is that the relation between analogy and grammaticalization (109-113) is based on a controversial definition of the latter ("a two-stage process consisting of reanalysis and extension" (110)). In fact, it is preferable to separate the two processes with reanalysis on one hand, and grammaticalization on the other. As Detges and Waltereit (2002: 152) have shown, these two kinds of changes "arise as a consequence of basic needs of communication, which can be identified as listeners' strategies in the case of reanalysis and as speakers' strategies in the case of grammaticalization". Secondly, there seems not to be a reanalysis-stage in every grammaticalization process (see Marchello-Nizia 2006: 45-46).

However, these criticisms should not deter from the interesting and stimulating ideas presented in this book. Itkonen's views of analogy make a real summa of this concept while opening a multitude of new epistemological and methodological perspectives for research in linguistics.

References

Detges, Ulrich and Waltereit, Richard. 2002. Grammaticalization vs. Reanalysis : A Semantic-Pragmatic Account of Functionnal Change in Grammar, Zeitschrift für Sprachwissenschaft, 21/2, 151-95.

Marchello-Nizia, Christiane. 2006. Grammaticalisation et changement linguistique, Bruxelles: De Boeck.

Monneret, Philippe. 2003. Le sens du signifiant. Implications linguistiques et cognitives de la motivation. Paris, Honoré Champion.

Monneret, Philippe. 2004. Essai de linguistique analogique. Dijon, ABELL.

Links

© 2002-present ICLA; all rights reserved.
Site admin information and contact