The book title indicates the main objective of the volume under review: to provide an introduction to the concept(s) of analogy and morpho-logical change with a special “focus […] on the domain where they intersect” (p. 1). Beyond that, the book “consider[s] the role that analogy […] plays in language history outside of morphology” (p. 2). For this purpose, Fertig critically examines conceptual and terminological fundamentals on the notions of analogy and analogical change (chapter 1). He first differentiates between analogy₁ and analogy₂. The first refers to the general cognitive ability to detect similarities in relationships between elements of one domain and to transfer these relationship patterns to other elements. The second refers to what is traditionally known as analogy: the process of building new forms on the basis of existing linguistic patterns. For example, the pattern think—thunk emerges by analogy with sink—sunk. In this example, thunk is what Fertig refers to as an analogical formation. The latter is not restricted to single words, but can also apply to phrases, clauses, sentences, etc. As analogical formations such as thunk are new and thus deviate from the norms of language use, they are also analogical innovations. At this point, the reader wonders if the terms analogical formation versus innovation are really necessary, as analogical formations always deviate from current norms before eventually prevailing in common use. If the latter indeed happens, an analogical change has occurred. Last but not least, Fertig subsumes the traditional notions of folk etymology and contamination under the term associative interference.

Chapter 2 mainly deals with different types of reanalysis. It introduces for the first time—at least to my knowledge—the labels as A-, B-, C-, D-reanalysis, which refer to the value/element that undergoes reanalysis in the equation A:B=C:D well-known from four-part analogy. In a C-reanalysis, for example, a C-form such as the former singular pease is reevaluated as a plural on the basis of relations such as beans (plural) : bean (singular), leading to the new singular form pea (p. 29). The B-reanalysis, on the contrary, results in a resegmentation of formerly separate elements. Compare, for example, the emergence of the German extended suffixes -ler and -ner, where the l/n was originally the stem-final consonant of words derived with -er, such as reden+er > Red+ner ‘speaker’ (p. 32).

A very interesting aspect discussed in the context of reanalysis is Fertig's observation that a verb's denominal status plays a key role in regularization and irregularization (§3.2.5). Thus the strengthening of originally weak verbs generally affects lexemes that have either lost their denominative basis or at least their semantic connection to the underlying noun/adjective (for example, German weisen ‘to show’, originally ‘to make wise’, derived from weise ‘wise’). Conversely, strong verbs “would have to be regularized if […] reanalyzed as denominal” (p. 30). This last prediction in particular needs to be verified by diachronic and crosslinguistic research. While most examples offered for the different reanalysis types seem to fit the picture, it is at least questionable whether the emergence of lexical suppletion such as go–went also relies on reanalysis (more precisely: D-reanalysis, §2.5.1). Regarding the relevant literature on lexical suppletion, such as Nübling Reference Nübling2000 and Ronneberger-Sibold Reference Ronneberger-Sibold, Boretzky, Enninger and Stolz1987, Reference Ronneberger-Sibold1988, its emergence seems to be much more complex to me (for go–went, see Bybee Reference Bybee1985:91–92).

Chapters 3 to 5 deal with various types of analogical change, considering first proportional (for example, the famous four-part analogy) and nonproportional change (that is, folk etymology, blends). Chapter 5, entitled “Problems and Puzzles”, addresses questions concerning analogical change, such as whether paradigm leveling is proportional or not (§5.1), as well as so-called analogical nonchange (§5.2) and phantom analogy (§5.3) phenomena. The latter refer to paradigmatic changes that appear to result from analogical leveling. Thus, the new infinitive New High German küren ‘choose’ (replacing Old High German kiosan ‘choose’) is (probably) not the result of leveling out the consonant stem alternation known as grammatischer Wechsel (Old High German kiosan, kor, koren). Fertig argues that küren is a denominal and thus historically weak verb derived from Kür ‘decision’ that merged somehow with the remaining paradigm of kiosan (p. 78). In contrast, analogical nonchange comprises phenomena such as Lass' Reference Lass1990exaptation (discussed in §2.5.6 in more detail), in which a formal distinction is retained, but it acquires a new function. For example, Indo-European aorist and perfect merge into a new past tense in Germanic by preserving the inherited ablaut distinction between aorist and perfect stems (p. 77).

It is commendable that the book does not confine itself to morphological change, but rather takes a detour to analogical change beyond morphology (chapter 6). It discusses syntactic (§6.2), lexical/semantic (§6.3), morphonological (§6.4), and phonological (§6.5) change, the interaction of analogy₂ and sound change, as well as questioning whether regular sound change can be interpreted as analogical (§6.6).

Chapter 7 addresses “[t]he question of how analogical change is constrained” (p. 120). Several universal and system-dependent accounts are taken into consideration such as Natural Morphology in the tradition of Mayerthaler's (Reference Mayerthaler1981) universal naturalness (§§7.3.3, 7.3.4) and Wurzel's (Reference Wurzel1984) system-dependent naturalness (§§7.3.5, 7.3.7), evolutionary grammatical theory (§7.3.8), as well as the role of token (§7.5) and type frequency (§7.3.6) in analogical change. Fertig comes to the conclusion that the “main principle that really seems to matter is that analogical change favors whatever patterns and pattern types happen to be most common in a language, in terms of type frequency” (p. 121). If type frequency is really the only thing that matters, phenomena such as Nübling's (Reference Nübling2000) Differenzierungsanalogien (p. 132)—that is, morphological changes/analogies that aim at increasing formal paradigmatic differences in extreme high-frequency words—remain unexplained. How can an originally weak verb such as German haben ‘have’ acquire the morphemic pattern called Wechselflexion of strong verbs? This pattern, which contrasts the 2^nd/3^rd person singular present from the remaining paradigm cells, is of low type frequency, both within German strong verbs (only 55 out of 170 types), and within German verbs in general (only 55 out of ca. 4,000 types).

In the context of the constraints on the intraparadigmatic direction of change discussed in §7.4, it would also have been important to mention Bybee's Reference Bybee1985 Relevance Principle of grammatical categories. Highly relevant grammatical categories such as tense for verbs or number for nouns are rarely affected by intraparadigmatic leveling compared to less relevant categories such as person/number or case. Thus, the 2nd person singular marking in the preterite of strong verbs realized through introflection (ablaut and umlaut) was leveled out, whereas the vowel alternations expressing tense were preserved, as in singen ‘sing’ : sang- ‘preterite stem’ versus süngest ‘2sg’»sang(e)st, but still singen–sang–gesungen ‘sing–sang–sung’ (see Nübling & Dammel Reference Nübling and Dammel2004).

In this regard, it is also important to mention that token frequency (§7.5) not only makes (irregular) forms more “resistant to analogical influences” (p. 118) but also has an influence on the direction of analogical change. As Bittner Reference Bittner1996, Dammel et al. Reference Dammel, Nowak and Schmuck2010, Dammel Reference Dammel2011, and Nowak Reference Nowak2015. have shown, single strong verbs do not shift by means of a single change but rather in a series of steps. Depending on the category frequency ratio between past tense and perfect, simple past forms can be weakened first, as in German (and Dutch), which undergoes perfect expansion: backen ‘bake’–buk>backte but gebacken still remaining strong (like English shaven, p. 2).

In Swedish, however, the past tense forms are used more frequently, which is reflected in verb class change. Here, past participles become weak first (see Dammel et al. Reference Dammel, Nowak and Schmuck2010). The situation is even more complex if the relevance concept is also taken into consideration. Strong class features that express less relevant verbal categories such as person and number are lost first, for example, NHG du bäckst, er/sie bäckt ‘you bake, (s)he bakes’>du backst, er/sie backt. The vowel changes expressing the highly relevant tense category (that is, past tense and past participle) are eliminated last. Thus, Fertig's assertion that “[e]ven the simple regularization [of strong verbs] usually appears to affect the past tense and the participle simultaneously” (p. 8) is—at least from a crosslinguistic perspective—untenable. The eighth and final chapter rounds off the volume with an interesting overview of morphological theory and recent psycholinguistic approaches such as exemplar-based versus rule-based models (§8.5) or the heated debate between single- (that is, analogical or rule-based) and dual-mechanism models (§8.6).

Overall, Fertig provides an insightful introduction to analogy and morphological change by critically analyzing, readjusting, and complementing traditional accounts on analogical change. It is particularly praiseworthy that Fertig offers many diachronic examples to explain and illustrate the different mechanisms of analogical and morphological change. The author also provides helpful summaries at the end of each chapter. Chapters 3–5, which are associated in content, are logically summarized together in §5.4. There is also a helpful index of introduced terms and cited linguists at the end of the volume (pp. 157–160). The volume itself is well edited and does not contain typographical errors. Only the book cover showing the Indo-European family tree does not perfectly fit the subject of the book, suggesting that the present volume is an introduction to language history or to Indo-European (languages). All told, I think that this book can also be recommended to students of historical linguistics at a basic level.