Since the term morphome was created by Mark Aronoff (Reference Aronoff1994) in his monograph Morphology by Itself, both theoretical and empirical studies have been conducted to demonstrate the autonomy of morphology. One example of a theoretical position on this is Paradigm Function Morphology (Stump Reference Stump2001), which holds that some phenomena can only be accounted for by morphology exclusively and morphology has its own internal architecture. Empirical studies that claim to provide evidence of the existence of morphomes include Maiden (Reference Maiden2018) on Romance verbs and Aronoff (Reference Aronoff1994) on English perfect participles. However, some issues, even those fundamental ones, such as the existence, especially cross-linguistic existence, and diagnosis of morphomes, remain controversial.
Herce’s (2023) monograph, The Typological Diversity of Morphomes: A Cross-Linguistic Study of Unnatural Morphology, is a significant and productive study in dealing with these morphome-related issues. It is often criticized that morphomes are defined negatively, therefore causing negative typological predictions (Koontz-Garboden Reference Koontz-Garboden, Luís and Bermúdez-Otero2016). Herce, on the contrary, adopts a positive definition of morphomes, which consists of three main concepts – syncretism, systematicity, and (un)naturalness. By separately clarifying the three concepts, Herce realizes the ambition of applying this definition in practice, which enables us to recognize morphomes in different languages using consistent criteria. As a consequence, a cross-linguistic morphome database is established as a main contribution of this study.
In Chapter 1, ‘Introduction’, Herce firstly remarks that this study aims to construct a cross-linguistic morphome database, considering that previous studies on morphomes normally concentrate on single language family, that is, Romance languages. The main object of this study, ‘morphome’, is firstly exemplified by ‘L-morphome’ and ‘N-morphome’ in Spanish. The first occurrence of the two terms, however, is potentially confusing for readers who are unfamiliar with Maiden’s work (e.g. Maiden Reference Maiden2018) where ‘L-morphome’ and ‘N-morphome’ are initially coined and substantially discussed. A similar terminological problem arises when Herce quotes the term ‘PYTA morphomes’ in Chapter 3. Nevertheless, readers can rely on other examples in the book to figure out what the ‘morphome’ is. A familiar example might be the English verb ‘be’. Forms in all plural cells and the form in the second person singular cell in the paradigm of this verb are the same, that is, ‘are’ and ‘were’. Later, Herce indicates that this morphome database is intended to answer questions on the diversity of morphomic structures and their synchronic and diachronic properties. With this purpose, a working definition of ‘morphome’ is therefore required. Herce adopts Trommer’s (Reference Trommer, Siddiqi and Harley2016: 60) definition of morphomes: A morphome is ‘a systematic morphological syncretism which does not define a (syntactically or semantically) natural class’. This means that the identification of morphomes is determined by three main concepts – syncretism, (un)naturalness, and systematicity, which are discussed extensively in Chapter 2.
Chapter 2, ‘Issues in morphome identification’, therefore, initially elucidates the nature of systematicity and unnaturalness and then deals with other issues that may arise in the identification of morphomes. Herce suggests that we can draw from morphosyntactic, diachronic, and allomorphic or morphophonological evidence to test the systematicity of a syncretism, in other words, to test whether formal identity exhibits a polysemous relation. The allomorphic or morphophonological evidence is suggested to be the best determiner of systematicity since it is more morphome-related than morphosyntactic evidence and more accessible than diachronic evidence. For instance, the English plural ‘-s’ and genitive ‘-s’ undergo the same range and distribution of allomorphs, that is, /s/, /z/, and /ɪz/, and are therefore classified as a morphomic pattern based on the allomorphic evidence. Herce then discusses the other fundamental concept – (un)naturalness – and proposes that there is a continuum between the most natural and most unnatural cases. Taking the biunique relation between form and meaning as the most natural case, morphomic cases that are unnatural can be found easily. Although these two concepts form the core of the definition, systematicity and unnaturalness are insufficient to screen out morphomic structures unmistakably. The following discussions, therefore, are devoted to raising the threshold for morphomicity to ensure the purity of the morphome database. In a nutshell, in addition to being systematic and unnatural, these additional requirements specify that most ideal morphomes are a) restricted in the inflectional paradigm of a single lexeme, b) based on exclusive forms that are overt and easily discriminated from neighboring phonological forms, c) not motivated by productive phonological processes, and d) instantiated by at least two different exponents. This chapter ends with a passing comment on morphomic cases in domains other than inflection, such as derivation and syntagmatics.
Chapter 3, ‘Morphomes in diachrony’, discusses how morphomes emerge, change, and disappear in a language. Morphomes may arise through one or more diachronic processes. Sound change is found to be the most important factor in the emergence of morphomes, although morphomes may differ from each other in terms of the results and locus of sound changes. Analogy, which can be motivated morphosyntactically or formally, is another main factor. As pointed out by Herce, the emergence of morphomes is more likely to be triggered by formally driven analogy. For instance, for Spanish verbs, the form of the second-person plural imperative is identical to the form of the infinitive, which is triggered by the historical formal identity between the two paradigm cells. Interestingly, such a kind of formally driven analogy can function across the affix-root boundary. In Spanish, for example, the suffixes of the third person plural past and gerund were the same, which prompts a stem identity between the two paradigm cells nowadays. As for the disappearance of morphomes, we can draw from sound change, analogy, the gradually decreasing productivity of a morphome, and the loss of morphosyntactic categories. Overall, this chapter reflects the cognitive reality of morphomes on the basis of their diachronic properties. Furthermore, cases of those formally driven morphomes reveal that compared to meanings, forms do play an equal or even more important role in the organization of inflectional paradigms.
Chapter 4, ‘Morphomes in synchrony’, is the core of this book, which can be divided into two parts. In the first part, Herce responds to two limitations of prior studies on morphomes: the absence of uniform measurement and the scarcity of morphome data. To overcome the first limitation, Herce provides specific ways to apply the two criteria of systematicity and unnaturalness, whose nature has been elucidated in Chapter 2. A third criterion is additionally generated to conciliate morphologists’ worry over incautiously regarding default forms as morphomes. To avoid this, any converse-type morphome whose cells constitute more than 70% of its paradigm cells is excluded from the database. By fulfilling the three criteria, morphomes collected in the database are most morphomic and situated at the end of the morpheme-morphome continuum. Even with this higher threshold for morphomicity, a cross-linguistic morphome database is constructed with 120 morphomes from 79 languages. Each morphome is described in detail in Section 4.2 with its synchronic and diachronic information. However, the diachronic changes of some morphomic structures are missing.
The other part of this chapter describes the morphome database quantitatively. These collected morphomic structures differ from each other formally and distributionally, which are embodied by their values on nine variables – degree of unnaturalness, degree of restrictedness by morphosyntactic features, word-form recurrence, paradigmatic recurrence, lexemic recurrence, number of exponents, the length of shared forms, informativity, and paradigm size. Principal component analysis is conducted over these values, from which Herce finds that Romance morphomes only represent a very small area of morphomes cross-linguistically, although they are always the main research object in the domain of morphome studies. Also, morphomic structures tend to be similar to each other in terms of the nine variables if they come from cognate languages. The remainder of Section 4.4 shows that the nine variables are correlated. However, this is contrary to the spirit of Multivariate Typology (Bickel Reference Bickel and Bril2010), which emphasizes an ideally independent relation among all variables/dimensions. Unfortunately, no further explanations are offered by Herce for this contradiction. Another unclear point is that Herce fails to illustrate the different roles of Canonical Typology (Corbett Reference Corbett, Frajzyngier, Hodges and Rood2005) and Multivariate Typology in the current study. Although both approaches define linguistic concepts from independent dimensions that are continual in nature, the logical reasoning behind them is different. Specifically, dimensions in Canonical Typology are developed deductively, whereas dimensions in Multivariate Typology are generated inductively. Also, in Canonical Typology, the ideal instance of a linguistic concept does not necessarily exist in reality, while the ideal instance identified in Multivariate Typology is the one that most frequently exists. As a consequence, linguistic concepts and their ideal instances that are identified from these two approaches may be different. Therefore, it is confusing and inappropriate to use a notion like ‘canonical morphome’ (224) under the methodology of Multivariate Typology. Nevertheless, the qualitative and quantitative analysis of the morphome database reflect that morphomes are not typologically unique, and some morphomic structures even appear in genetically unrelated languages. It is also found that less unnatural morphomes tend to be more frequent, which is unexpected in light of the characterization of morphomes as unnatural.
Based on the findings of this study, Chapter 5, ‘Implications: Features and forms’, reexamines two accepted views in the field of morphology. One view that morphomes are characterized as being independent from morphosyntactic features and values is challenged since it is found that the distribution of morphomic structures is conditioned by certain morphosyntactic features and values (Sections 4.3.1 & 4.3.8). Also, some morphomes are generated by morphosyntactically driven analogy, and morphosyntactically motivated morphomes are more frequent than those morphosyntactically unmotivated ones. The other view that meanings are superior to forms in morphology is criticized. This study provides evidence that, similar to meanings, forms can equally influence paradigmatic distribution because there is an implicative relation between different forms, both paradigmatically and syntagmatically. Paradigmatically, the distribution of a paradigm may be a template for the distribution of another paradigm. Syntagmatically, one form may always co-exist with another form, the occurrence of one form, therefore, predicting the existence of the other. In consequence, both forms and meanings contribute to our understanding of morphology, no matter natural or unnatural.
This monograph ends in Chapter 6, ‘Conclusions’, where Herce summarizes the main content in each chapter, and the findings in Chapters 3 and 4 are reiterated further to emphasize the innovations and significance of this study. Finally, Herce suggests that experimental research should be conducted on morphomes to corroborate the findings in this study.
To conclude, this book deserves a recommendation. Firstly, it concentrates on an unnatural but indispensable part of inflectional morphology – morphomes. This reveals that inflectional morphology is actually composed by both natural and unnatural paradigmatic distributions, and more importantly, the naturalness is a continuum. This study also has certain implications for other morphological domains that may have unnatural phenomena, such as derivation and syntagmatic distributions. These are promising research directions in the future. Lastly, the methodological part of this book provides detailed guidelines for linguistic typologists who aim to combine quantitative analysis with typological studies.
