Björn Wiemer & Juana I. Marín-Arrese (eds.), Evidential marking in European languages: Toward a unitary comparative account. Berlin & Boston: De Gruyter, 2022. Pp. xvii + 732.

Evidentiality research has generally been focused on grammatical markers encoding source of information (e.g. Aikhenvald 2004). For ‘non-morphological’ evidential marking, however, there have not so far been any comprehensive and in-depth surveys, even for very well-described languages (e.g. English, German, French, Russian). To close this gap, the volume unites a collection of language-specific profiles of ‘extra-grammatical’ evidential expressions in Germanic, Romance, Slavic, and other languages, with an attempt to create a unified account for units with evidential core meanings in European languages.

This collection begins with an introduction in which the editors specify the conceptual premises crucial to the characterization of evidentiality. Following some of Anderson’s criteria (1986), the range of evidential units are defined as those ‘conventionalized markers of discursively secondary status that scope over propositions’ (9) with inherent evidential functions. Hence, the assumed lexicon-grammar cline (Wiemer and Stathi 2010) is slightly modified as the basis for distinguishing grammatical and extra-grammatical evidential markers. Structurally autonomous forms with core evidential meanings are also subsumed into the heterogenous inventory of extra-grammatical marking of evidentiality, including sentence adverbs, function words (particles and complementizers), predicatives (e.g. uninflected units of the Polish widać ‘can be seen’), constructions (derived from SEE-verbs, SEEM-verbs, and SAY-verbs), even modal auxiliaries and adpositions. Following Marín-Arrese (2015), contributors in this volume restrict their focus to Indirect-Inferential (IIE) and Indirect Reportative (IRE) as two subcategories of evidentiality.

The next four parts of the collection probe into empirical investigations focusing on ‘extra-grammatical’ evidential marking in specific language profiles, drawing on evidence from large general corpora The British National Corpus (BNC), Corpus of Contemporary American (COCA) and those with a more specialized focus. Statistical results from corpora indicate both intriguing recurrent phenomena concerning marker types, functional patterns, etc., and phenomena that are more saliently represented in some languages while being less frequent in other languages.

Parts II and III concentrate on the language profiles in Germanic (English, Dutch, German) and Romance languages (French, Spanish, Galician, Catalan, Portuguese), respectively. Both language groups share a similar inventory of dedicated indirect evidential markers in combination with a divergent system of multifunctional units. Typical IIE and IRE markers include verbal expressions and constructions associated with a set of cross-linguistically recurrent types of hyperlexemes with a prominent evidential function: SEEM-verbs, SEE-verbs, SAY-verbs, the necessity modal MUST and ACCORDING TO-phrases, together with certain indirect-inferential adverbials. Inferential evidentiality is predominantly encoded by markers derived from inferential sentence adverbs (e.g. apparently in English, kennelijk, augenscheinlich in Dutch and German, aparentemente in Spanish and Galician, visiblement in French); SEEM-verbs (English appear, seem, Dutch lijken, blijkens, schijnen, German scheinen, French paraître, sembler, Spanish and Portuguese parecer, etc.); and MUST-verb constructions (e.g. must+INF in English; moeten, müssen+INF in Dutch and German), together with particular constructions with SEE-verbs in English and Spanish. For IRE expressions, there are a number of clear-cut hearsay markers consisting of reportative adverbials (e.g. allegedly, reportedly in English, soi-disant, prétendument in French, het schijnt dat, naar verluidt in Dutch); constructions associated with reportative modals (e.g. zou+INF in Dutch; sollen and wollen+INF in German); and SAY-verbs and HEAR-verbs (e.g. dit-on in French, se dice and he oído+que in Spanish, dir in Catalan), and some ad-positional phrases denoting evidence from the third parties, (e.g. according to in English; zufolge, laut, nach in German; au dire de/aux dires de in French).

With regard to the multifunctional units, however, Germanic and Romance languages appear to be slightly different in semantic extensions between IIE, IRE, and modality. On the one hand, semantic extensions between IIE and IRE pertain to particular verbal expressions with constructional variants especially for modals and SEEM-verbs. For example, inferential extensions of modals are particularly associated with infinitival constructions of MUST-verbs in both language groups and some quasi-modals in Spanish (amenazar+INF ‘threaten’ and prometer+INF ‘promise’). In a similar vein, reportative extensions from inference are closely related to the high constructional variability of SEEM-verbs. For example, the infinitival constructions (e.g. seem+INF in English; schijnen+te-INF in Dutch) evoke either inference or hearsay, while the reportative readings occur most consistently in the impersonal constructions (e.g. het schijnt dat ‘it seems that’ in Dutch; parece ser (que) ‘it seems to be (that)’in Spanish; sembla/pareix que ‘it seems’ in Catalan). Frequent reportative expressions are also found with impersonal and parenthetical constructional variants of BELIEVE and SEE-verb constructions in Spanish and Catalan (e.g. se cree ‘it is believed’; por lo visto ‘from that seen’).

On the other hand, such semantic extensions seem to become conventionalized to a lesser extent in certain Germanic languages than in others. For example, inferential extensions from modality ‘seem to have developed to a somewhat lesser extent in German than in the closely related language Dutch’ (138). German scheinen is mainly inferential in its evidential use, whereas its Dutch counterpart schijnen functions either as an indirect-inferential evidential or as a clear-cut reportive. The same holds for the German modal müssen, which has not developed a hearsay meaning, while the Dutch modal moeten can be used with both inferential and reportive readings. In addition, constructions associated with SEE-verbs and verbs of cognitive attitude in English are predominantly used for expressing IIE, while reportative meanings are associated with the impersonal constructional variants of SEE and BELIEVE-verbs in Spanish and Catalan (e.g. por lo visto ‘from that seen’; es veu que ‘one sees that’; se cree ‘it is believed’). Moreover, SAY-verbs in English are clear-cut IRE markers, while IIE can be marked by French SAY-verbs in the conditional mood (e.g. dirait-on and on dirait ‘looks like, seems’), among others.

Parts IV and V move on to the exploration of non-morphological means of indirect evidentiality in Slavic (Russian and Polish) and other languages (Lithuanian, Greek, Basque). Evidential expressions corresponding to the abovementioned cross-linguistic hyperlexemes ‘SEEM, MUST, SAY, SEE, ACCORDING TO, etc.’ are also identified in these language profiles as IIE and IRE markers. Basque displays the same evidential hyperlexemes as identified for the Germanic and Romance languages. On the other hand, Slavic languages, together with Lithuanian and Modern Greek, demonstrate some different characteristics regarding the word class of evidential markers, functional preference on the type of evidence, and semantic extensions from inference to hearsay.

For the marker types, predicative is a salient word class in Russian and Polish for encoding evidentiality, and most evidential predicatives are heterosemic units. That is, the range of their syntactic behavior is not restricted to predicate function: they always behave like particles as well. For example, poxože ‘similarly’ and kažetsja ‘seem’ in Russian and widać ‘see’, słychać ‘hear’, znać ‘know’ in Polish are used either as a predicative or as a particle. As for functional preference, the authors note that the inferential zone seems to be the richest evidential subdomain in Russian, while very few markers can encode reportive meanings. Regarding semantic extensions, only perception-based inferentials with a medium degree of epistemic support have been observed to acquire a reportive function in Russian. Therefore, the subclassification of inferential units in Russian can be distinguished by parameters, including: perception/non-perception evidence; degree of epistemic support; and presence/lack of contextual support, the distinctions of which may contribute to a more fine-grained grid of criteria for other languages. In comparison, verb-based, noun-based, adjective-based, and prepositional markers are predominantly used in Lithuanian for expressing evidentiality, while the major categories of evidentials in (Standard) Modern Greek derive from grammatical morphemes, function words, and constructions with speech act verbs.

The concluding section in Part VI presents a synthesis of the ‘evidential profiles’, the convergent and divergent properties of evidential units, and implications from a cross-linguistic perspective.

In sum, this collection gives a comprehensive survey of the ‘evidentiality profiles’ of 13 European languages, focusing on distinct markers and construction types that were neglected in previous studies. Evidentiality is treated on a par with domains like temporality, actionality, and modality and with grammatical categories like tense, aspect, and mood (TAM paradigms) as a background. The account of evidentiality marking beyond obligatory grammatical means makes a substantial contribution to an adequate characterization and delimitation of evidentiality as a conceptual domain.

In addition, a large set of common hyperlexemes with prominent indirect evidential readings are identified in European languages ranging from verbal constructions, modals, adverbials to prepositions, and such recurrent types of evidential constructions may serve as the basis for comparison for further cross-linguistic analysis of evidential marking across languages both in and out of Europe.

Last but not least, the findings in this volume echo the views and claims on evidentiality from various branches (e.g. cognitive-functional linguistics, typology, contrastive linguistics) concerning functional-orientation, multifunctionality, and form-meaning correlation (e.g. Langacker 2017; Mortelmans 2017; Foolen 2018). For example, the polysemy of SEEM-verbs, which has been discussed in Germanic and Romance languages (e.g. Vliegen 2011; Mortelmans 2017), is greatly elaborated on, with a survey of its function ranges and iconic relation with personal involvement (Chapters 2, 7, and 15, etc.). The findings in this volume are not isolated achievements but represent a progression of systematic research on evidentiality over the years.

All in all, this volume assembles extensive and language-specific lexicography of evidential markers in European languages towards a comparable basis. By doing so, it contributes significantly to obtaining ‘a panoramic picture of the complexity of the evidential strategies’ (Liu and Wang 2020) and to illuminating our understanding of the omnipresent Tense-Aspect-Modality and Evidentiality (TAME) categories. Other than the abovementioned merits, the interactions between the subdomains within evidentiality (between grammatical and extra-grammatical evidentials; between direct and indirect evidentials, etc.) are not much addressed, which might preclude a more thorough understanding of the complexity of this intriguing phenomenon. However, with fruitful empirical evidence and explicit demonstrations, this volume is valuable for specialists interested in evidential marking from all backgrounds, even for those with no specialties in contrastive or typological linguistics.