Abstract

• I. Mapping the Field: Ambiguity and the Changing Landscape of Literary Evaluation
• II. Tracing Ambiguity: Four Case Studies
• III Adding Ambiguity: Summary and Outlook
• Bibliography

I. Mapping the Field: Ambiguity and the Changing Landscape of Literary Evaluation

[1]Today, millions of people read and evaluate literary books and stories online.‍[1] These practices are embedded within a digital, networked culture that hinges on a common technological infrastructure.‍[2] Since the onset of global digital transformation in the late 1990s, ›users‹ have become prototypical social actors,‍[3] using social media platforms, apps, and other digital devices.‍[4] Among the various online practices related to consumption, news, and social interaction, it is striking how frequently literature becomes a subject of engagement – through awarding stars and likes, writing reviews, producing and sharing videos (on platforms such as TikTok and YouTube), as well as commenting on blogs and websites or even writing fan fiction.‍[5] On the one hand, these individuals act as consumers sharing their personal experience with the purchase (›fast delivery‹) and the product itself (›the book was so good it made me cry‹); on the other hand, they increasingly function as intermediaries within the literary domain, and even as gatekeepers who »actively participate in constructing the worth of the objects they review« – a role traditionally performed by literary critics.‍[6]

[2]Social actors as prosumers,‍[7] or wreaders,‍[8] have assumed a dual role in a cultural field undergoing profound transformation. But why are they so numerous? And what motivates them? A possible starting point lies in the human need for stories. As Gerhard Lauer puts it, we are »hungry for stories and will probably continue to be for a long time«.‍[9] Storytelling – the sharing of interpretations of and guidelines for everyday experience, serving to reduce contingency and complexity – appears to be one of the key practices of humanity, apparently a universal trait.‍[10] The stories we continue to tell have remained remarkably constant over millennia: they are »still about love and trust, heroic courage and fear, betrayal and other social messages«.‍[11] And, interestingly, despite most recent substantial media changes the book as modern cultural artifact for storytelling has retained – or even regained – its popularity (see Feldkamp et al., Freudenau et al., this volume).‍[12] ›Book love‹, to use a current buzz word, denotes the attachment to books as physical objects, their iconic status as culturally valued goods, and the act of reading itself.‍[13]

[3]In this volume, however, we are not so much interested in the stories or the material literary artifacts themselves, but in people’s perspectives on them: book love and the need for stories seem to be coupled with a need for interaction about books and stories. The evaluative discourse of readers about aesthetic artifacts, including literature and books, is »surely as old as art itself,«‍[14] yet under changing medial, economic, and social conditions. Every day, people now engage in millions of digitally afforded evaluative practices revolving around ›the book‹ and their (shared) experience of it.‍[15] Their practices are driven by aesthetic, affective, and social motives: in his recent book Wut und Wertung‍[16], Johannes Franzen argues that (especially negative) emotions towards aesthetic artifacts are a key driver of cultural change. This argument is backed by Rita Felski’s and Ika Willis’s observation that affect permeates literary reception on all levels.‍[17] Regarding the social dimension, over a decade ago, Andrew Piper pointed out that a need for sharing is at the center of digital media used for reading: »We want other people to read the same thing we are reading (commonality); we want to be able to send other people what we are reading (transferability); and we want to be able to talk to other people about what we are reading (sociability)«.‍[18] The social dimension of online reading has been taken up by Federico Pianzola, who maps out a research paradigm of »Digital Social Reading«, combining a detailed taxonomy of online evaluation and reception practices with large-scale quantitative analyses that have become possible because of the datafication and platformization of social reading.‍[19]

[4]The actions that communities of practice perform around literary artifacts are typically evaluative and thus connected to values – and value systems – that are explicitly or implicitly invoked. If we look at Goodreads, the largest English-language reviewing platform, launched nearly two decades ago, a reviewer’s assertion such as »What a great book, I thought«‍[20] can potentially refer to style, content, their interplay, but also to material aspects such as cover design and paper quality, or to emotional responses and intertextual positioning – i.e., how the book relates to other books within its genre or by the same author. In this case, we are dealing with a review of Jenny Erpenbeck’s novel Kairos, and the sentence refers to the structural dimension of the novel, to its intertwining of two different types of content – a love story and a political history, with the latter being pitched at an allegorical level of the prior:

[5]»I really love how Erpenbeck primary [sic!] writes a love story, with the description of the life in DDR in the background. She gives a great picture of a piece of history, without actually writing the history.«‍[21]

[6]With Renate von Heydebrand and Simone Winko we understand literary evaluation as an implicit or explicit linguistic action, through which an object of reference is assigned a particular attributive value that is based on an axiological value.‍[22] In the example above, the reviewer’s evaluation is personal, and so far, positive: the reviewer evaluates the object of reference, the book, explicitly and positively ( »I love how Erpenbeck primary writes a love story«; »a great picture of a piece of history«, emphasis ours) and harnesses as axiological basis of their evaluation both the structural form of the novel as well as the particular choice of content. So far, the evaluation is unambiguous, since an »attributive value is clearly assigned to a reference object in the linguistic-pragmatic context at hand and no other readings or scope for interpretation questioning the evaluation can be identified« (Moskvina / Petzold, this volume, our translation). Yet, after this opening paragraph, the reviewer goes on to show what they did not like about the book:

[7]»At the same time, the book was a disappointment. It is way too long and the story looses [sic] its magic. Over the middle hundred pages, we get several repetitions of the male protagonists [sic] jealousy. It is tiresome and it should have been revised with a sharp knife. This part makes my admiration for the author drop. I get it, it is a picture of an immature and egocentric man. Still, it is a way too big part of the novel.«‍[23]

[8]In a way, the review is thus ambiguous. At the same time, it is structured very logically, and in fact, transmits a clear and unanimous message: one argument (›great because of structure and content‹) is contrasted with another one (›disappointment, because too long and repetitious‹), and together they form a type of synthesis in a final statement: »The book is absolutely worthy of reading. But it started out so much better than it turned out«. In accordance with their judgment, using the affordances of the platform Goodreads, the reviewer awarded three out of five stars.

[9]This example represents a type of reviewing which so far has not been in the focus of literary or cultural sociological studies as such: it assigns no clear condemnation (or »fail«), but is no clear song of praise either; the assigned rating sits in the middle section of the ordinal scale. In online reviewing, ambiguities come in many shapes. The above example is an almost ideal case of a dialectic evaluation that proceeds in an impartial, yet engaged way (concluding that the book is a worthwhile read, albeit with clear shortcomings). Although there is no overall positive or overall negative judgment, taking it all together, the reviewer’s assessment is in fact not ambiguous. At the same time, as we will see, vagueness and context dependency give rise to other forms of ambiguity – ones in which it becomes difficult or even impossible to determine how an evaluation is to be interpreted. Indeed, indeterminacy and underspecification are not the exception, but rather the norm, as Anna Moskvina and Kristina Petzold (this volume) observe.

[10]Reviewing ambiguity may also emerge through what has been termed ›ostentatious subjectivity‹, a feature characteristic of many online reviews. As Renate Giacomuzzi argues, subjectivity plays a major role in the »symbolic currency« of online lay criticism: authenticity.‍[24] Authenticity is closely tied to the value of reviewers’ autonomy from cultural hegemonies and from the market, as a direct, and untainted grasp. At the same time, it is a legitimization for emotional ad hoc evaluations, and importantly, for contradictory judgments. Here is where authenticity can foster additional forms of ambiguity – including the expression of conflicting values and moods and the diplomatic communication of negativity (see also Spengler, this volume).

[11]The contributions in this volume are thus interested in how different types of ambiguities manifest in digital reviewing practices: in natural language use (Freudenau et al., Spengler, Moskvina / Petzold), in the numerical aggregations on the platform (Feldkamp et al., this volume), and in the interactions between the two (Moskvina / Petzold, this volume) as well as between reviews and other types of quality judgments (Feldkamp et al., Spengler, this volume).

[12]When applying a wide definition of the arts as aesthetic and fictional practices around artifacts that people experience in everyday practices,‍[25] i.e., the texts that are in fact being read (or listened to), bought, downloaded, and discussed – fictional texts of all types and at all levels of difficulty, including neglected ones, longsellers, or bestsellers‍[26] come into view. From here, the next step is to ask questions about what those many, many people ›on the web‹ actually do when they interact with fictional and literary texts evaluatively. This incorporates the role of uncertainty and ambiguity in today’s world that is shaped by a proliferation of choices and where a palpable contingency‍[27] clearly (and maybe even especially) affects the cultural and aesthetic sector, and thus literary experiences. Assuming that people normally apply strategies to reduce uncertainty,‍[28] one might expect online lay reviewing and rating to be largely unambiguous – especially when considering the medial affordances of online social media discourse that appear to foster short and seemingly clear statements.

[13]This assumption is strengthened when we think about the ›metric‹ dimensions of today’s society that Steffen Mau has also called »evaluation society«:‍[29] a constant and massive practice of (online) reviewing, facilitated by digital medial affordances, gradually shifting important power relations. For example, in the cultural sphere, evaluation culture has put the traditional hegemony of the ›high brow culture‹ and ›professional criticism‹ under pressure, not only by ›shit storms‹, ›review bombs‹, and the like, but by the quantity and popularity of repeated, networked, practice on social media. As a result, Johannes Franzen argues, have matters of taste become matters of an extended public debate, with many more (types of) participants than earlier: a much higher diversity in tastes and values is now forcefully publicly articulated. For the historical structural change of media and communication, Jeff Jarvis suggests an extension of Marshall McLuhan’s bodily metaphor: »[N]ow that all may speak, we have moved from the age of the ear (orality) to the eye (text) and now to the mouth (networked discourse)«.‍[30] Amateurs have claimed an equal voice, and while the hegemony of a bourgeois ›high culture‹ has eroded, clashes between high brow and mass culture have become more frequent as well as more fervent.‍[31] Part of the amateur’s emancipation is a new type of cultural critical professionalism that is not primarily based on school certificates, degrees, and job contract (the literary critic), but on popularity, credibility, and merits accumulated within platformed communities of practice and remuneration by ›the book market‹ (the influencer, the super user). If measured in these terms, certain types of ›lay reviewers‹ of the web are thus in fact professionals (see Spengler, this volume). As Renate Giacomuzzi argues, the reviewers-prosumers‍[32] are much more successful multipliers for the publishing houses than the writers of the ›feuilleton‹, the culture pages, that do not even exist any more in many newspapers across the globe. The distinction between lay and expert reviewers thus appears increasingly untenable: »The attribute of professionality is directly related to the connection with the market.«‍[33]

[14]An important aspect of the networked, transmedial, attention-oriented nature of digital communication is that it not only enables direct contact between formerly separated socio-cultural spheres, but that it also rewards and prompts affective participation. Both dimensions together seem to make online practices of evaluation typically favor ›strong‹ types of evaluative statements, or as Andreas Reckwitz puts it, a »digital affect culture of the extremes«.‍[34] Research thus suggests that networked digital evaluation is typically polarized: acts of affirmation and acclaim on the one hand, and »dark«, that is hateful and destructive »participation«‍[35] on the other. In the context of online lay book reviewing, this may lead us to expect either ›fails‹ or ›songs of praise‹.‍[36] But yet, there is a considerable number of lay reviews and evaluations out there that are in fact not straightforward, polarized, or single-minded. This is precisely where this special issue is directed:

• Are such online reviewers behaving like the traditional literary critics that ponder pros and cons?
• And aren’t reviews rather clear (and thus non-ambiguous) if they can pinpoint strength as well as weaknesses of a book?
• Are the truly ambiguous reviews thus those that are indecisive, unable to take a stance?
• Or those who do not clarify their criteria for judgment?
• Are other reviews ambiguous because negative judgment is very subtly addressed in superficially positive, sugarcoated reviewing?
• How unambiguous are thus seemingly clear reviews with four or five star ratings?
• And are some of those apparent ambiguities even artifacts generated by a numerical dimension of the discourse, where average ratings falsely suggest ambiguity, while different sub groups of reviewers actually pull into opposite directions?

[15]It appears remarkable that readers still remain »the most elusive objects of literary study«.‍[37] After the reader reception studies shift in literary studies in the 1970s, which however put an emphasis on aspects of the text,‍[38] with some exceptions,‍[39] literary studies have not asked many questions about readers. But this seems to be changing. Since the 2010s, the literary-cultural landscape has co-evolved with the digital transformation and produced an increase of literary participation which has been seen as a democratization, for example by Gerhard Lauer‍[40] and/or a decline of critical professionalism as voiced among others by Rónán McDonald,‍[41] Sigrid Löffler,‍[42] and Moritz Baßler.‍[43]

[16]Yet, despite the availability of abundant and diverse data for studying evaluative reading practices, reviewing practices are still of rather marginal importance for literary studies. This might be because the new forms of ›doing literature reviews‹ are embedded in a participatory culture and foster democratization of cultural evaluation both at the level of the actors and that of the artifact. The wreaders of the net are normally not trained critics, and the books they write and talk about are typically ›popular‹ ones. But ever since ›everyone is a critic‹,‍[44] at least those scholars who work on the subject matters of contemporary literature and forms of literary criticism have started to pay more attention to the phenomena of popularization, participation and polarization in the literary field, and thus, ›the reader‹. Under these new conditions, new types of questions are being asked about canonicity and cultural as well as aesthetic Bildung in a society that lends more visibility to style-communities.‍[45]

[17]Forms of ›hedonistic‹ reception, once scolded under the ›art paradigm‹, are now increasingly licensed – including practices such as skipping sections, abandoning books mid-read, or labeling canonical works as »boring«.‍[46] By comparison with some twenty years ago, a new quantitative diversity in the basic values underlying literary reviewing has become manifest: high-brow taste, with its embrace of the canonical (as cultural heritage), its inclination towards aesthetic challenges for elaborated pleasure, and its orientation on a Kantian impartiality and autonomy (see Spengler, this volume), is now but one way of approaching literary artifacts among many. The cultural sphere as a whole has become more popularized – openly ›hedonistic‹ tastes appreciate entertainment, affect-orientation and identification, as well as ›simple‹ recognizable schemata and material books as objects of fandom and sub culture. The digital revolution has thus leveled certain power relations (and produced new ones). The effects of those substantial (and ongoing) changes still have to be fully understood, as do the aspects of doing literature reviews that in fact continue with earlier practices.‍[47]

[18]As reported, a growing body of research located at the intersection of literary studies, digital humanities, media studies (computer mediated discourse) and cultural sociology has shown how readers interact, evaluate, rank, recommend, and thereby produce ›reviewing culture‹ themselves.‍[48] Here, the metric, or numerical, dimension of online reviewing by stars or hearts arguably plays an important role: numbers convey non-ambiguity, simplification, verifiability and neutrality – and as Johannes Spengler puts it, by means of reduction, quantification appears to produce order.‍[49] But does an ordinal middle position actually tame uncertainty, or objectify it? We do not yet know, as most studies have focused on evaluative extremes, particularly positive evaluations.‍[50] Another question is whether books can even be meaningfully compared on the basis of metrics and ratings: doesn’t the tertium comparationis (the common denominator of the comparison) vary from genre to genre and type of audience? And is it maybe unclear in the first place? (see Moskvina / Petzold, this volume). Such questions have hardly been asked, as ›ambiguous‹ dimension of evaluations, including the use of ordinal middle positions (›three out of five stars‹) of rating scales, but also textual reviews that juxtapose both positive and negative aspects of a work, and truly ambivalent, uncertain, or plainly ambiguous statements, have, to date, received little attention by research.

[19]This is all the more surprising given that non-straightforward evaluations and reviews can mean a whole range of things: They may depict a differentiated weighing up of the weaknesses and strengths of the evaluated text. Else, they may be identified not predominantly by mid-scale ratings, but in the failure to provide enough context or reasoning for one’s attributive values – in the text. In some of these cases, but not in all, mid-scale ratings may arguably reveal problems with aesthetic and moral judgment, or an insecure grasp of value systems as posited by cultural criticism. In others, even positive ratings with five stars may be quite ambiguous. And in large parts of the book reviewing sphere, the topic of negative judgment is closely related to forms of ambiguous reviewing, as there seems to be a tendency to embed negativity within a friendly, polite message. By contrast to other online spheres, this sphere runs on the principle of amiability, following from the literary critic’s traditional motto »Punch Up, Never Down«.‍[51] Within such horizontal, peer-oriented communities of practice, ›playing it nice‹ becomes a coping mechanism for navigating epistemic, social, and affective uncertainties – a dynamic that minimizes face-threatening acts (see Freudenau et al., this volume, Moskvina / Petzold, this volume, Spengler, this volume). As for instance Philippa Chong’s empirical cultural-sociological study demonstrates, even contemporary forms of ›professional‹ literary reviewing often reflect similar constraints.‍[52]

[20]In any case, the sheer existence of non-straightforward reviews is at odds with the scholarly and medial attention to polarization, impermeable positivity biases, and unleashed negativity in today’s networked digital society. It appears that a closer look at ambiguous and negative evaluations‍[53] is promising for an exploration of readers’ and reviewers’ actual evaluative and appreciative behavior, potentially prompting an update of theories of the social-cultural sphere.‍[54]

[21]Such an endeavor, however, poses methodological challenges. Much online reviewing can be analyzed either at large scale or presents itself by aggregated measures such as numbers of likes or average ratings. On large platforms such as Amazon, Goodreads, or LovelyBooks, the average ratings must not to be taken at face value: mean average ratings‍[55] over many books, such as in between 2.5 and 3.5 stars on a scale ranging for instance from 1 to 5 stars, might in fact mask variation across the actual individual ratings, with potential positions towards the extremes of the scale. It is thus necessary to find study designs and methodological approaches that discern ›really‹ ambiguous, indecisive, or pondering evaluations from numerical artifacts (see Feldkamp et al., this volume, for an approach that factors in a number of platform-independent proxies for literary quality). A question that is imminent for platformed reviewing is the role of metrics such as Likert scales or the number of likes, especially in relation to the textual reviewing practices: how do ordinal and numerical judgments relate to evaluative language with its content and sentiment dimensions?‍[56] Anna Moskvina and Kristina Petzold (this volume) take up this question and ask whether the positive reviews as indicated by four and five star ratings (and the like) on online reading platforms are necessarily unambiguous. This approach allows them to assess in a differentiated way ›clarity‹ (unambiguity) of judgment, potentially independent on the polarity of the evaluation (positive, negative, neutral).

[22]The papers in this volume ask about different aspects of literary evaluation on the web that impact the way in which readers judge books – among these are different genres and types of objects,‍[57] types of (intended) audiences, different groups of reviewers, and the specific platforms and their uses.‍[58]

[23]Also, as mentioned above, the axiological values driving and emerging from the practices of literary reviewing are of interest: what are the (literary, book-related) premises brought forward by the user when judging the quality of types of books in ambiguous, indecisive or ambivalent ways? What role do the material features of the book play? How is textual beauty and style as a formal feature addressed? How about offers for affect and identification? Is the book situated within a relational system of other books, and genres, maybe also literary periods?‍[59] In the case of children’s literature, for example, how do reviewers balance didactic-educational values with literary-aesthetic dimensions? How do reviewers address picture books as a medium that combines multiple sign systems (language and pictures) and that forms a complex multi-modal or intermedial product? What impact does the reception in reading-out loud situations play – especially considering that the audience consists of both adults and children (Cf. Freudenau et al. , this volume for a case study of a children’s picture book)?

[24]Finally, any inquiry into online lay reviewing must address the linguistic and medial expressions of value systems underpinning literary evaluation. A driving factor of doing the reviewing is authenticity, which appears to license content ambiguity and moderates a language of politeness (which through indirectness fosters a semantic ambiguity). Another one is attention that in turn drives originality. Here, many questions remain to be asked in a fast changing sphere about the ambiguities of linguistic practices at the levels of whole reviews, as well as sentences, phrases, adjectives and other linguistic signs, such as emoji font types, and color, but also other modes such as facial expressions, tone of voice, gestures, and dance, as well as the digital affordances of the platform such as the ordinal scales of stars and likes, and their aggregated averages. This volume is just the beginning of asking such questions, and hopefully will prompt many more.

II. Tracing Ambiguity: Four Case Studies

[25]The special volume has its origins in a panel at the 27th Germanist*innentag, whose insights and findings are deepened and expanded by the contributions collected here. Its aim is to shed light on the phenomenon of ambiguous (and negative) literary evaluation under the conditions of the digital transformation of literary criticism. To this end, we have collected studies reporting concrete case studies across genres and languages that depart from different methodological and theoretical premises.

[26]The four papers included in this special issue all look at cases of ambiguity in practices of online literary reviewing. Literary reviewing is a cultural practice mediated by a number for factors, such as the affordances and constraints of their specific medial form, group-related conventions, the object of review, and the stable or situation-dependent dispositions of the user and their self-concept (see Moskvina / Petzold, this volume). As our papers are situated in different academic traditions, they vary considerably by main perspectives assumed, as well as by the (genres of the) reviewed books, the books’ intended audiences, the types of digital media and online platforms (Goodreads, Amazon, BücherTreff.de, book blogs, Instagram), the review communities, their digitally represented collective practices, and the typical aspects of the reviews in terms of reviewing form, content, and (underlying) values.

[27]The contribution by Pascale Feldkamp, Yuri Bizzoni, Mia Jacobsen, Mads Rosendahl Thomsen, and Kristoffer L. Nielbo is a computational literary study with a strong quantitative component, motivated by questions from the sociology of aesthetic success and reader appreciation. It addresses aspects of the numerical in people’s practices on Goodreads, a highly popular and thus statistically large book review platform. The authors approach the quantitative data available from reviewers’ contributions on the platform as a proxy for literary success and reader appreciation: by a rule of thumb, higher average ratings indicate a greater success, and higher average appreciation. Such extrapolations are possible because the platform aggregates many readers’ opinions indicated by ordinal ratings into one quantitative value: the average rating. Previous studies have typically focused on predicting literary success by analyzing titles considered the ›very best‹ or ›very worst‹. In contrast, and in alignment with this volume’s interest in ambiguous and potentially ›mediocre‹ reviewing judgments, the paper is interested in books with average ratings accumulated from ratings in the middle section of the ordinal Likert scale that ranges from 1 to 5 stars.

[28]Interestingly, the paper shows that such average platform ratings do not necessarily indicate ›mediocrity‹ as a collective judgment. Rather, average ratings might in fact obscure a polarized rating behavior. To interrogate the nuanced nature of »mediocre« ratings, the authors conducted an empirical analysis, departing from a dataset drawn from the Chicago corpus (N = 9,000 novels published in the United States between 1880 and 2000) to control for distribution and popularity (books included in library holdings around the globe). From this corpus, they extracted a subset of novels (n = 2,150) that occupy the middle quartile of Goodreads average ratings, specifically those falling within a range of 3.72 to 3.91 on the Goodreads rating scale. To control for platform effects, they identified ›successful‹, ›prestigious‹, or ›canonical‹ works within this subset by additional proxies of literary appreciation. The analysis revealed that the books with average ratings at platform scale can be taxonomically classified into three distinct subgroups: (1) Books that the majority of readers rated in fact as ›neither great nor terrible‹, corresponding to a number in the ›middle‹ of the ordinal scale (2 or 3 stars per unique review) on the platform Goodreads, and that in addition did not feature as ›high quality‹ by any of the other literary quality proxies consulted; (2) books that again receive mostly ratings in the ›middle‹ of the ordinal scale on the platform Goodreads, but that are at odds with other proxies of literary quality. These in fact seem to polarize a wider evaluative landscape in literary culture. Those books are typically less well-known than the books in the last group. (3) Books that polarize audiences on Goodreads as well as in other proxies of literary quality. Those are typically widely read (e.g., Nabokov’s Lolita) and considered to be ›great‹ by a portion of Goodreads users, but as ›terrible‹ by another portion: Their apparent mediocrity is an effect of two very decisive readerships pulling into opposite directions. The polarization effect is reflected also by the other literary proxies.

[29]The contribution by Tanja Freudenau, Marlene Antonia Illies, Jan-Niklas Meier, Ulrike Preußer, Sandra Siewert, and Christian Volkmann puts a main focus on educational and didactic dimensions of literary evaluation and puts center stage a genre that is marginalized in (traditional) media: children’s books. Here, the net, through large reviewing platforms as well as other sites, in fact works as a motor of emancipation in Jeff Jarvis’s sense,‍[60] giving those books and their reviewers ›a voice‹. In particular, the contribution is interested in »negative and ambiguous-reflective evaluation practices on a sales-oriented online platform«.‍[61] The paper thus addresses the topics of (1) multimodal or intermedial narrative text, (2) parents / educators of pre-school children as an intended, or factual, audience, and (3) questions of underlying axioms relating to morality and aesthetic quality. The authors depart from the observation that children’s book reviewers typically focus on straightforward pedagogical-didactic criteria rather than on those of aesthetics and literary quality. By contrast to other genres, the review genre typically incorporates a full summary of content (no spoiler alert necessary), an assessment of readers’ age, tips for conveying the predominant educational purpose by means of pedagogical actions, and a usually clear purchase recommendation. If presented at all, literary aesthetics judgments are normally not explicitly rooted in assessing the text’s details. Before this background, the authors adopt a position that criticizes the predominant educational ›instrumental focus‹ of children’s literature criticism. Of particular interest are the comparatively fewer number of reviews that contain negative or ambivalent evaluations. These potentially require a more detailed justification and deliberation of composition and (aesthetic) effect. The paper presents the results from a computer-assisted analysis of a corpus of negative or ambivalent reviews of Marcus Pfister’s The Rainbow Fish (German: Der Regenbogenfisch, 1992), a popular contemporary classic of children’s literature that has accumulated almost 2,900 reviews on German-language Amazon from 2002 to 2023. The quite naïve morals of the story address the fish’s essential corporality, its rainbow scales, which invest it with extraordinary beauty, but which is met by jealousy by the community and apparently coupled with the fish’s arrogance. Aspects like the simple social-psychological mechanics and a potentially unreflected degree of violence have provoked criticism, yet, on the online platforms the positive rating supersede the negative and mediocre ones by far, with an average rating of 4,7 out of five stars. The paper examines the 3.8 % of all collected reviews that allocate either one, two, or three stars. In this dataset, using an annotation procedure for literary axioms slightly adapting Heydebrandt / Winko 1996, the authors find that ›morality‹ is overall the most frequent axiom, and also the value that is most often coupled with a negative evaluation. They also find a strong qualitative relation between morality and an interpretative form of recounting the story, where textual passages cater to justify the negative evaluation of content. Rather than a rebuttal, this is a specification of Peter Boot’s‍[62] tentative observation that positive evaluations more often contain a recounting of the story: By contrast to the full content accounts of the story observed by Boot, the recounting here is strongly evaluative and apparently selective. Furthermore, the authors find that reference to children is often made and thus the reviewers appear as gatekeepers: they recommend / do not recommend the book for children. In addition, by contrast to other genres of book reviews, the children’s book reviews do not normally put a focus on the axiom of subjective pleasure and the flow of reading experience, but on the didactic merits of the read – an axiom of effect-oriented action guidance (key word: ›suitability for children‹). Furthermore, the multimodality of the picture book appears as a salient feature and is overall positively related to the axiom ›beauty‹ – but many reviewers do not give any reasons for their judgment although a discussion of aesthetic criteria may be fathomable.

[30]Finally, from their study, the authors suggest two types of evaluative ambiguity: global ambiguity if reviewers do offer arguments for evaluative statements, but do not arrive at one clear overall judgment; and local ambiguity where isolated evaluative statements are made without justifying arguments, remaining superficial statements of taste due to a lack of reflected premises.

[31]The paper by Anna Moskvina and Kristina Petzold presents a mixed-methods analysis of online reviews of books of different literary and non-literary genres that puts a focus on types of ambiguities of digital evaluation practices. From an interdisciplinary perspective in between literary studies and corpus linguistics, the authors look at the linguistic forms of evaluative practices and ask to what extent ›clarity‹ in the linguistic expression can function as a possible distinguishing criterion between evaluation practices on different platforms and between ordinal and textual evaluations. Their hypothesis is that the specific usage environment of platforms such as German-language Amazon that are more strongly oriented towards quantification (by awarding stars) and direct purchasing decisions produces more unambiguous reviews. In contrast, non-commercial platforms centered on exchange and discussion, where quantification is optional (such as BücherTreff.de), appear more conducive to ambiguous or nuanced reviews. The authors analyze a review corpus using three different methods: manual annotation, dictionary-based sentiment analysis, and qualitative content analysis. Surprisingly, they find no statistically significant differences across platforms for the categories ›clearly positive‹, ›clearly negative‹, ›ambiguous‹ and ›neutral‹.‍[63] What emerges instead is a general dominance of positive reviews relative to ambivalent or argumentative ones, and a striking rarity of clearly negative reviews. This positivity bias may result from pre-selection effects: users often choose books they expect to like, meaning negative reviews typically occur only after erroneous purchases. While negative comments do occur, they are more frequently found at the sub-review level rather than as overall judgments. The authors thus argue that ambiguous reviews are the primary place for criticism both on Amazon and BücherTreff.de: either users only rarely find books truly awful (see above), or they just do not judge ›bad‹ books clearly negatively. The authors also discuss that predominant purchase platforms such as Amazon are approached as spaces of recommendation, and which are thus not modeled on traditional literary-critical discourse, where books may well be judged as utter failures.‍[64] Rather, one here expects the transmission of potentially successful decisions of purchase and reading.‍[65]

[32]At the same time, the two corpora vary most distinctly in the proportion of evaluative statements in general: Amazon reviews contain more evaluative statements than BücherTreff.de, where much more recounting of the story and book metadata happens – an observation that the authors explain with the predominant function of the platforms.

[33]However, in both corpora they find the same kind of significant difference for the distribution of ›clarity‹ across star categories – the proportion of non-ambiguous reviews is highest for five-stars reviews and lowest for three-stars reviews. The ordinal middle position thus most often contains positive and negative statements at the same time, as well as ambivalent statements.

[34]On a methodological note, they find a difference between manual annotation and sentiment analysis, which may indicate that style in online reviews is often figurative (judging from the failure of sentiment analysis to disambiguate figurative language and negation).

[35]Furthermore, the authors examine the relation between ordinal ratings and different types of ambiguity, finding that linguistic communication often does not translate into the rating: the statements contain more ambiguities than the ratings may suggest. Here, interestingly, weighing of axioms by individuals or with reference to communities plays a role, as well as unintended inconsistencies in terms and concepts and conditional limitations of evaluations. Finally, there are books that are viewed as deliberately triggering ambiguous or ambivalent reviews and that lead some reviewers to high ratings especially because of this effect (see also Feldkamp et al.). The authors conclude that the widespread assumption of a digital review culture dominated by »praise and foolishness« – i.e., under-complex, unequivocally positive lay reviews – cannot be sustained. Although clearly positive reviews are common on Amazon and BücherTreff.de, their study demonstrates that the clarity of literary evaluation depends on far more variables than the platform alone or the presumed critical competence of the reviewing public.

[36]In the final paper, Johannes Spengler addresses the topic of negative judgments in the book blogging sphere. From the vantage point of applied literary studies, he takes a deep look into the connections between contemporary ›lay‹ online criticism and the history of literary criticism. As key aspects he discusses the ideal of an impartial and autonomous criticism confronting the commercial practices of the book market, as well as amateur professionalism under the conditions of the digital transformation. He departs from the observation that in today’s book market, bloggers and book influencers have been taking on the role of professional intermediaries. He argues that in literary criticism generally, the genre of the polemic negative review (German: ›Verriss‹) has been seen as an antidote to dependency on the book market. With the advent of digitization, from around 2000 onwards, this role appears to have become even more urgent, as evaluation practices are increasingly embedded in an attention economy: the lines between reviewing, popularization, and advertisement are becoming even more blurred. From this perspective, Spengler observes that the fundamental dilemma of literary criticism seems to repeat itself in the field of amateur critics: asserting one’s own impartiality and autonomy. His essay examines which strategies amateur critics use to voice negative criticism, and how book bloggers deal with negative judgments. Drawing on user-generated content from blogs, Instagram, TikTok, and YouTube, Spengler finds that amateur critics often employ discursive methods of evaluation similar to those of professional critics. With reference to Phillipa Chong’s work, he argues that negativity is largely attuned by politeness strategies, and the motto »Punch Up, Never Down«. While politeness is a universal mechanism for dealing with uncertainties of epistemic and social types, in today’s many-to-many communication of the networked online literary reviewing communities, it is crucially mediated by the symbolic currency ›authenticity‹: users appear honest and unbiased while communicating with the target audience, their peers, at eye level. A pivotal feature of online lay reviewing is that the authority of judgment is not primarily constructed any more by argumentation in the textual reviewing practices (as it was / is in professional criticism), but by field dynamics: the distribution of visibility and attention – which in turn is administered by the seemingly objective ratings and rankings of the ›Like Economy‹.‍[66] As to ambiguity and uncertainty, it is interesting that specifically the clarity of the judgment (both in praise and in reprimand) is the grounds on which traditional highly prominent figures like Marcel Reich-Ranicki based the reputation or credibility of the criticism. By comparison, the face-saving, the uncertainty-taming strategy ›playing it nice‹ of lay critics of the net when addressing shortcomings and negative judgments overall begets more ambiguity of a certain type. In conclusion, Spengler argues that the forms of community that are emerging on the net give evidence of a high degree of self-reflection on evaluation practices and strategies for dealing with instrumentalization by literary producers.

III Adding Ambiguity: Summary and Outlook

[37]This special issue aims to add a largely missing perspective to the research on online lay literary evaluation: ambiguity. In the course of our editorial work – alongside key methodological and data-related questions – we became increasingly aware of the wide range of forms of ambiguity that emerged across the contributions. This raised questions not only about the aggregation of ratings on platforms, but also about the textual dimensions of particular reviews, as single sentences can be quite clear in isolation, yet produce ambiguity when embedded in larger textual contexts. We also learned about the role of politeness and ambiguity especially for negative reviews and the role played by the current ›communication paradigm‹ of authenticity. We touched upon the issue of context, missing information, and need for disambiguation at word, sentence, and utterance level. Interestingly, ambiguity reveals a dual nature: in today’s complex, information-saturated and diversified world, it can signal a breakdown of sense and values – but it can also serve as a strategy for producing meaning and fostering sociality across different communities (and platforms). We observed that ambiguity in evaluative practices varies between communities, but also across genres – for example, in children’s books, where underlying premises as a rule monitor ethical deliberations, but not questions of aesthetic pleasure. When asking about the change, or constancy, of ambiguous online lay reviews by comparison to traditions of professional literary criticism, it became clear that autonomy of taste and judgment are still an ardent issue, however with almost reverse perspectives on ›the popular‹ and ›the elevated‹. We learned about adjectives and other linguistic signs used in ambiguous and unambiguous ways, which is a suitable ground for further studies that look at the forms of ambiguity in evaluation at the linguistic and multimodal levels, including videos, body-movement and tonal pitch. Clearly, culture is ever evolving, fueled by people’s need for (fictional and non-fictional) stories and thought, as well as needs for sociality, and affect. In a multi-voiced and encompassing network of expressions of tastes, pleasures, and opinions, ambiguity – in its many facets – appears to play a much more important role than previously assumed.