Criticism and politeness strategies in academic review discourse : a contrastive ( English-Italian ) corpus-based analysis

Drawing on a corpus-based approach, this paper explores the mitigation strategies used to soften criticism in English and Italian book review articles in the disciplinary field of linguistics. Most corpus-based analyses on academic criticism have focused on the use and function of politeness strategies in English academic review genres, mainly book reviews. Recently, an increasing number of studies on academic book reviews have examined the issue from a cross-cultural perspective. This study attempts to contribute to the area of cross-cultural research on reviewing practices by exploring how criticisms are managed in a somewhat neglected review genre in academic discourse studies – the book review article. Criticisms will be identified on the basis of their lexico-grammatical features and further categorized into “direct” and “mitigated” (Itakura & Tsui 2011, 1369). The mitigation strategies identified in both language corpora mainly involve the use of sequences of speech acts such as praise-criticism, criticism-praise, criticism-suggestion, praise-suggestion, and hedging. However, their distributions reveal differences in the two languages. While praise is prominently used in both English and Italian book review articles, Italian-speaking linguistics reviewers employ a lower proportion of hedges than their English-speaking colleagues and are more likely to opt for suggestions as a form of indirect criticism. The results demonstrate that linguistics reviewers writing in English and Italian deploy a considerable range of linguistic devices when expressing mitigated criticism of peers. Their use and distribution are discussed in relation to national/cultural writing conventions, but also differences between “large” and “small” disciplinary cultures (Holliday 1999). Some implications for EAP learners and practitioners are also considered.


Introduction
It is widely recognized that academic discourse is intrinsically linked to ways of establishing and disseminating knowledge (e.g.Flowerdew 2002;Bamford & Bondi 2005;Hyland & Bondi 2006).Scientists belonging to academic discourse communities are driven by both the desire and the need to communicate new knowledge to the other members of the academic community by means of publication.Within the academic context, reviewing plays a crucial role in the process of knowledge dissemination and evaluation of others' research, while providing an alternative forum in which academics can set out their views (Hyland 2000).This suggests that reviewing is highly interactively loaded as judgements carry significant social consequences (Hyland & Diani 2009).It may involve authors in making counterclaims (i.e. they offer alternative claims to those made in the literature they review), and/or indicating a knowledge gap with reference to previously published work (Swales 1990).These rhetorical options may involve the criticism of members of one's own discourse community.This issue has been the centre of intense scholarly interest in the area of academic communication, focusing on the interpersonal pragmatic dimension commonly referred to as "professional disagreement" or "academic criticism/conflict" (e.g.Myers 1989;Hunston 1993;Hyland 1998).These studies on criticism in the research article genre have suggested that criticism of previous texts is usually conveyed indirectly, with frequent hedging strategies intended to minimize potential face-threatening acts.This contrasts with other academic genres where criticism is expressed more explicitly, e.g.blurbs, book reviews, book review articles, literature reviews (e.g.Motta-Roth 1998;Hyland 2000;Hyland & Diani 2009;Salager-Meyer & Lewin 2011), medical journal editorials (Salager-Meyer 2001;Giannoni 2011), or the 'conflict article' (Hunston 2005), an article where a writer explicitly opposes a named other writer or paper.This suggests that the expression of criticism may vary in "the context of particular genres" (Paltridge 1997;Hyland 2000).
Criticism is a potentially face-threatening speech act, and so appropriate politeness strategies are very often employed to compensate for any face-threat incurred (Taylor & Cameron 1987;Heritage 1989).Criticisms are seen as threatening the addressee's (H's) "positive-face-want" as the act of disagreement (or contradictions, challenges) indicates that the speaker "thinks H is wrong or misguided or unreasonable about some issue, such wrongness being associated with disapproval" and therefore implying "(potentially) that the speaker does not care about the addressee's feelings, wants, etc." (Brown & Levinson 1987, 66).
Most analyses of academic criticism based on data taken from English-speaking academic communities have drawn on the frameworks of politeness theory.For instance, Myers (1989) sees passives and pronoun use in research articles as strategic politeness devices.Similarly, Johnson (1992) and Johnson and Roen (1992) find that compliments in students' peer reviews are used to mitigate face-threatening acts.In his study on praise and criticism in book reviews, Hyland (2000) shows that praise is often used to open an English academic book review as an effective way to mitigate subsequent criticisms.He also points out that praise is frequently used to close a book review so as to protect the positive face of the book author by repairing the negative effects of the review writer's earlier criticisms.In each of these studies, praise and criticism are seen as enactments of strategic politeness, drawing on Brown and Levinson's model of face-maintenance.Their face model proposes that it is usually in everyone's best interest to satisfy the face needs of interlocutors and, as Hyland (2000, 45) rightly notes, "criticisms are particularly risky as they undermine a hearer's positive face, the desire to be approved of and have one's goals seen as desirable".
The politeness-based approach has also been effectively applied to the study of praise and criticism in a highly conventionalised written genre that is overtly evaluative in nature, the book review.This has been done from a cross-cultural perspective to shed light on the realisation of praise and criticism across different academic communities.For example, a contrastive study of negative evaluation in medical book reviews in English, French and Spanish by Salager-Meyer and Alcaraz Ariza (2004) has shown that criticisms are mitigated more frequently in Spanish and French than in English.Similarly, Itakara and Tsui (2011) in their study on criticisms in academic book reviews in English and Japanese have found that Japanese book reviewers are more indirect than their English colleagues.On the other hand, by analysing the expression of praise and criticism in English and Italian economics book reviews, Giannoni (2006) has observed that in both English and Italian book reviews negative evaluation tends to be expressed implicitly.Lorés-Sanz (2012) interestingly found that in Spanish book reviews negative evaluation is almost nonexistent.
In this study I attempt to contribute to the area of cross-cultural research by focusing on a somewhat neglected review genre in academic discourse studies -the book review article. 1 The genre offers a critical analysis of a book (that is not necessarily a new publication) and presents a reassessment of the book's specific topics in the context of the debate surrounding the issue: a debate that the reviewer constructs not only with the reviewed author but also within the disciplinary area.As with all review genres (e.g.book reviews, review articles, book blurbs, literature reviews), the judgment of 'good' 1 See Diani (2012) for a more detailed discussion of the book review article genre.
and 'bad' is a central feature of the genre under examination, but this is presented as debate to construct a dialogue with the reviewed book author and other voices in the community.This debate allows the reviewer to create a 'research space' for his or her own views, exploiting the reviewed authors' reported opinions to construct a 'niche' for his or her claims on the topic (Swales 1990).
The focus of this paper is on the expression of criticism as it seems more problematic to find an appropriate way to comment negatively than to praise.The aim is to explore how criticism is managed in book review articles taken from linguistics journals published in English and Italian.The study therefore discusses the use of mitigation strategies for reducing the impoliteness or face-threat of a critical comment.
The context of this study is provided by a previous work (Diani 2015), which looked at the mitigation strategies used to soften criticism in English and Italian book review articles in the field of history.The aim of the present study is to extend previous observations to the field of linguistics, with the intent of exploring disciplinary and academic traditions.
The paper begins with a brief description of the book review article genre (Section 2).The data and methodology used in the present study are outlined in Section 3. Section 4 discusses the use of criticism and the way it is managed in English and Italian academic book review articles respectively.The article concludes with a discussion of findings (Section 5).

The book review article
The book review article (BRA) is a review genre that is something of a hybrid.It often appears in the book review section of academic journals, yet differs from a book review (BR) and a review article (RA).While a BR focuses on a recent book or cluster of related books and a RA surveys on a specific issue or area of research, a BRA combines critical evaluation of a book (not necessarily a recent one) with an extensive discussion of the issues raised with a view to supporting a specific position.The reviewer is always present: while BRs typically move from description to evaluation, BRAs also use critical statements to argue a position not only with the reviewed book author but also within the disciplinary area.
One identifying feature of the BRA is reporting the ideas an author discusses in his or her book as a point of departure for the reviewer's development of an evaluative discourse, as shown in (1) below, taken from the English sub-corpus under examination (cf.Section 3.1).
(1) At the end of chapter 3, which introduces the Chomskyan view of language acquisition as 'growth' of biological entities called 'grammars', Lightfoot concludes that language is an epiphenomenon, and that 'the notion of a language is not likely to have much importance if our biological perspective is taken' ( 74 2) below, where the reviewer's purpose is signalled and the argumentative structure indicated.The Critique, representing the body of the article, is usually characterized by a highly cyclical structure: each single issue identified is discussed critically in a series of sections.
(2) The narrative structure of my discussion is organized as follows: in section 2, I consider a representative number of Postal's core arguments and their factual basis in some detail, and argue that, contrary to his substantive claims, there is nothing like congruence between the B-extractions on the one hand and the environments which tolerate attested RPs on the other.In section 3, I review a number of his assertions about the class of conjuncts which, following Lakoff (1986), Postal considers in connection with CSC violations and argue that here too, his proposal fails to take into account the full range of facts.In particular, I show that these conjuncts fail his own criteria for the identification of weak (in his terms, selective) islands, and that, in particular, there are good cases of extraction from antipronominal contexts within these conjuncts.

The study
Cross-cultural studies on the expression of praise and criticism in Italian and English review genres are still relatively rare, and often limited to book reviews (Giannoni 2006).The analysis carried out here focuses on a less conventionalized review genre -the book review article.Cross-linguistic examination of how criticisms are managed in the genre under investigation will pay particular attention to the values reflected and implied by the different strategies used in the two languages.The general questions addressed here are: 1. Which types of mitigation strategies are used as a means of softening the threat of a critical comment in English and Italian linguistics book review articles?
2. What kind of lexical items are used as hedging with criticism in English and Italian book review articles?
3. Are there any cultural variations in the way criticism is managed by English-and Italian-speaking communities?

Data
The study is based on the analysis of two small comparable sub-corpora of book review articles taken from linguistics journals published in English and Italian respectively.Their composition was conditioned by the fact that the book review article is a genre which is not published regularly in academic journals.For each language, I consulted the most representative linguistics journals as indicated by disciplinary experts.Only a few journals were found to publish book review articles fairly regularly.From these, I took all the articles available spanning the years 1999-2001.Both sub-corpora are between 190,000 and 200,000 words.As regards the choice of the languages considered for comparison, no attempt was made to restrict the choice of reviewers to native speakers because the aim of the analysis was to describe what gets published in well-established journals of the English-and Italianspeaking community of linguists.Only review texts that were explicitly presented as translations were excluded.The two sub-corpora used for this study should thus be taken to be representative of writing for the Italian national community and writing for the international (English-speaking) community.

Methodology
The methodology combines tools from discourse analysis and corpus linguistics.
Discourse analysis contributes to the definition of pragmatic functions of mitigated criticism, whereas corpus linguistics offers ways of looking at language items.
The search for instances of criticism was checked against quantitative data by means of the wordlist function of WordSmith Tools (Scott 2008) Concordances of all the critical expressions identified were compiled and filtered manually.This manual filtering process was necessary to make sure that only occurrences of the negative expressions in which a critical comment was made on the book under review were included.
The criticisms identified in the sub-corpora were categorised into "direct" and "mitigated", borrowing Itakura andTsui's (2011, 1369) terminology: "mitigated criticisms were negative evaluations accompanied by linguistic devices which serve to minimize facethreatening effects, while direct criticisms were those that were not mitigated."Direct criticisms are not avoided in the corpora, and are in fact an integral feature of the genre, but they are often mitigated in various ways.Five mitigation strategies were identified in the sub-corpora: praise-criticism, criticism-praise, criticism-suggestion, praisesuggestion, and lastly hedging.The speech act sequence praise-criticism and hedging were drawn from Hyland's (2000) study of criticism in book reviews.A 'praise-criticism' pair is a criticism prefaced by praise that serves to soften the negativity of the criticism (Hyland 2000, 55).An example from the English sub-corpus is "Despite the deserved praise, we have some doubts about the coherence of the pedagogical strategy adopted by the volume".
Hedging has been approached in different ways in the literature, and there are differences between scholars as to the forms and functions that hedging is taken to assume. 3In this study, hedging is approached in broad terms as a rhetorical strategy used by reviewers to mitigate the interpersonal damage of critical comments, through the use of linguistic items conveying epistemic modality (Hyland 2000, 56). 4 4 Results

Mitigated criticism in English linguistics book review articles
To trace expressions of mitigated criticism in the English data, a careful manual sorting of the concordance lines of the critical expressions identified in the English sub-corpus was necessary.By carrying out this analytical step, 354 instances of criticism were identified.Of them, 204 (58%) were direct or without any mitigation strategies, while 150 (42%) were mitigated.In order to work towards a picture of mitigated criticism, the concordances of the 354 instances of criticism were compiled and filtered manually.As a result, five types of mitigation strategies have been identified.Their realisation and distribution in the sub-corpus are shown in Table 1.The use of the term goes back to Lakoff (1972) for whom 'hedges' were devices like rather, very, or sort of that can make an utterance fuzzier or less fuzzy.Since then the concept has been widened and defined differently by discourse analysts and pragmaticists.On the development of the notion see Markkanen and Schröder (1997) and Hyland (1998).
A closer look at Table 1 shows interesting quantitative results.If we consider the frequencies of each type of the mitigation strategies found in the English sub-corpus, criticism accompanied by praise figures highest overall with 84 occurrences (56%), followed by hedging which constitutes 34%.Suggestions are also used as a form of indirect criticism, combining with either praise or criticism, although their frequency is low (15 occurrences -10%).

Praise-criticism and criticism-praise
More than half of all mitigated criticisms in the English sub-corpus were accompanied by praise, either in the form of 'praise-criticism' (45%) or 'criticism-praise' (11%), the adjacency of the two acts serving to create a more balanced comment, slightly softening the negativity of the overall evaluation, as shown in ( 3) and (4) below.
(3) There is much historical discussion in this book that is interesting and it contains a rich compilation of valuable quotes and examples.However, it should be clear that I find Muhlhausler' main arguments largely unconvincing.(EN_LIBRA) (4) I rather regret that the final word is, in one sense at least, given to Honey's agenda in a serious book of this sort, even in rebuttal (cf. Trudgill 1998).But it is interesting that Crowley takes Honey to task, I think rightly, for invoking the same criterion of 'educatedness' that Trudgill does.(EN_LIBRA) In (3) the reviewer syntactically subordinates criticism to praise by prefacing a negative comment (it should be clear that I find Muhlhausler's main arguments largely unconvincing) with a positive one (there is much historical discussion in this book that is interesting and it contains a rich compilation of valuable quotes and examples).In (4), on the contrary, the reviewer first presents his negative evaluation (I rather regret) and then introduces his positive comment (it is interesting).The linguistic feature that typically characterises these sequences is the use of adversative connectors such as nevertheless, yet, although.These connectors indicate that the reviewer's opinion is not one-sided and that s/he can give some consideration to both sides of the topic discussed.

Criticism-suggestion and praise-suggestion
Though not very frequent, a further strategy identified in the sub-corpus involved 'criticism-suggestion' pairs.Criticisms combined with suggestions seem to expand what might be seen as a blunt criticism, developing what was lacking in the reviewed book.
The suggestion of an area needing development is typically realised as a hypothetical clause which includes the modal conditional verb would and could (in many cases, followed by a perfective: have+past perfect).The concordance of this pattern is shown in (5) below.
(5) EN_LIBRA concordance of would and could in the criticism-suggestion pair little more direction (theoretically and conceptually) would help identify and locate some of such extreme circumstances (Dorian 1981: 155).It would have been of greater value for the front pages of Anglo-Saxon journals.One would also have liked to see frequencies of word processor are mentioned).This section would have benefited from work like the lack of broader support for independence.She would have done well to explain, in by parsing considerations' (249).More interesting would have been the demonstration ut of context.The editors at Routledge could have made GL much more A few lines selected from the concordance, showing more context than in (5), are repeated in ( 6) and ( 7).
(6) This particular argument does not work because nobody denies that useful things may have inconvenient side-effects in some contexts <…> Lightfoot admits that the ECP 'may well be functionally motivated, possibly by parsing considerations' ( 249).More interesting would have been the demonstration that a UG principle is completely or primarily dysfunctional.(EN_LIBRA) (7) While this is the case, it nevertheless leaves obscure the contradictions between so much Puerto Ricanness among Puerto Ricans of different ideologies and the lack of broader support for independence.She would have done well to explain, in some detail, the complexities that characterize what to some seems contradictory about this position, while to many it does not.(EN_LIBRA) The full force of the criticism (this particular argument does not work; it leaves obscure) is assuaged by a suggestion (more interesting would have been; she would have done well).Since telling one's peers the advisable course of action constitutes a strong claim to knowledge, a recourse to a strategy which conveys a recommendation indirectly is invariably judged to be better.
Suggestions not only occurred in a 'criticism-suggestion' pair.Perhaps surprisingly, they also occurred in statements of praise, as shown in ( 8).
(8) While the speech community framework may be adequate for this analysis, Fr.Colman's discussion would have been further enriched by a nonlinear approach that accommodates "nests" and "networks" in complex speech communities (Kerswill 1994) such as Ireland.(EN_LIBRA) It seems here that the 'praise-suggestion' pair is being used less to compliment the reviewed author on an aspect of her/his piece of work than to call attention to some weakness.Once again, the suggestion (Fr.Colman's discussion would have been further enriched) works for the reviewer as an indirect criticism, signalling a problem in a way which is less threatening to the reviewer-reviewed author relationship.

Hedging
As presenting, denying and evaluating claims in academic writing constitute facethreatening acts (Myers 1989), the use of hedges repairs such potential threats. 5Not surprisingly, in such an evaluative genre, hedges are extensively employed to tone down criticisms.They account for 34% of all mitigation strategies.Table 2 shows the hedges used in the sub-corpus.As shown in the table, the linguistic forms used as hedges in the sub-corpus are those expressing epistemic modality, particularly those pertaining to possibility.By using hedges, reviewers refrain from making straightforward assertions.For example, the modal verbs may in ( 9) and might in (10) can be taken to express the kind of tentativeness intended to hedge criticism puts forth.Mauranen also discusses hedging in academic writing and describes academic discourse as "a world <…> where it is natural to cultivate hedges" (1997,115).

No. of occurrences
Similarly, the epistemic stance adverbials possibly and probably are employed as ways of mitigating the interpersonal damage of the reviewer's speech act of criticizing, as shown in ( 11) and ( 12). ( 11) Kirby might possibly be guilty of this in one or two places.(EN_LIBRA) (12) This is an indication that he probably lacks the appropriate reference data.(EN_ LIBRA) Reviewers also seek to hedge the effect of their criticism by using the modal lexical verb seem, as the concordances in (13) illustrate.
(13) EN_LIBRA concordance of seem* (including the forms seem, seems, seemed) rce historiographical wars, and Buel and Newell seem, at face value, to contradict one eneous time of chronology.Said then seems to confuse representation (in the to a general incapacity in learning (p.37) seems to me to be incorrect.of which speakers have tacit knowledge.Lightfoot seems unable or unwilling to conceive of those I consulted, however, the reflexive in (5) seems quite questionable or unacceptabl unique in music (pp. 183-184).This argument seems somewhat unfounded, since is guilty of poor field techniques.Muhlhausler seems to be completely unaware of how As we can see, the concordance is full of negative criticism that reviewers make less forceful through the modal lexical verb seem, used to weaken the threat of disagreement realised by verbal criticism (e.g.contradict, confuse), or adjectival criticism (e.g.incorrect, unable, unwilling, questionable, unacceptable, unfounded, unaware).
In addition to the traditional tokens of epistemic modality, however, the scope of hedging is here extended to involve other linguistic items.For example, the adverbial somewhat is frequently used in the sub-corpus to reduce the force of the reviewer's critique, as the concordances in (14) illustrate: (14) EN_LIBRA concordance of somewhat As noted above, this notion may be somewhat counterintuitive at first, Rhyme and reason resolves somewhat unsatisfactorily.In part s appealed to in this context is often somewhat crude and rhetorical, us ar participles like gefragt also seems somewhat incongruous, given that the of Indonesia, the inclusion of which is somewhat puzzling given its lack of accounts' seems to be overly broad and somewhat outdated.For example All the linguistic expressions examined above can be said to realise Brown and Levinson's strategy of being conventionally indirect in critical scholarly writing.

Mitigated criticism in Italian linguistics book review articles
As in the English sub-corpus, the search for instances of mitigated criticism in Italian started from the corpus data.An automatic search was followed by manual filtering of the concordances of the critical expressions identified in the wordlist of the sub-corpus to see what kinds of comments were made about the reviewed book.As a result, 68 instances of criticism were identified.If we compare this number with the instances of criticism found in the English sub-corpus (354), we can speculate that Italian-speaking linguistics reviewers are far less inclined to criticise their peers than their Englishspeaking colleagues are.This tendency was also found in the way criticisms are managed in both sub-corpora.In the Italian sub-corpus 35% of all criticisms were direct, while 65% were mitigated.By comparison, in the English sub-corpus, 58% were direct while 42% were mitigated.
In terms of type of mitigation strategies, however, the two languages show similar realisations.The way mitigations of criticism were realised and distributed in the Italian sub-corpus is shown in Table 3.Like the English sub-corpus, half of all mitigated criticisms were accompanied by praise, either in the form of 'praise-criticism' (39%) or 'criticism-praise' (14%), as ( 15) and ( 16) illustrate: (15) I fenomeni di adattamento fonetico e morfologico sono dettagliatamente indagati anche attraverso il confronto tra i due sistemi linguistici; resta invece in secondo piano l'analisi della componente semantica <…>.(IT_LIBRA) 'A very detailed investigation on phonetic and morphological phenomena has been carried out through a comparison of the two linguistic systems; however, the analysis of the semantic component remains a little blurred <…>' ( 16) Nonostante questa carenza, nel primo capitolo vengono dati esaurienti criteri e strumenti per la valutazione delle letture presenti nei manuali per l'insegnamento, di cui viene fornito in bibliografia un elenco piuttosto aggiornato.(IT_LIBRA) 'Despite this shortcoming, the first chapter contains comprehensive criteria and tools for assessing the readings contained in the textbooks, a rather updated list of which is provided in the bibliography.' In ( 15) praise (interessante 'interesting'; dettagliatamente indagati 'a very detailed investigation') is syntactically subordinated to a criticism so as to mitigate the damage of the critical comment (resta invece in secondo piano l'analisi della componente semantica 'however, the analysis of the semantic component remains a little blurred').In ( 16), on the contrary, negative evaluation comes first (Nonostante questa carenza 'Despite this shortcoming) and then a positive comment is introduced (nel primo capitolo vengono dati esaurienti criteri 'the first chapter contains comprehensive criteria'), but they serve the same mitigating purpose.
The second most frequent mitigation strategy employed in the Italian sub-corpus was the use of suggestions in criticism-suggestion pairs, amounting to 27% of all mitigation strategies identified.As we have seen, this strategy was also present in the English subcorpus, although it accounted for only 6%.
( The findings show that the amount of criticism is five times greater in English book review articles: in normalised terms per 100,000 words, 177 vs 34 acts.This tendency is also confirmed by the general preference for English-speaking reviewers to use direct criticism (58%) whereas Italian-speaking reviewers show a clear preference for mitigated criticism (65%).The results differ considerably from Hyland's (2000) study of hedged and unhedged criticism in English book reviews, in which 65% of critical expressions were mitigated.This might relate to differences in the primary functions of the two genres.The length of book review articles allows more opportunities for reviewers to develop an argument and engage with readers than the shorter book review.Reviewers exploit the reviewed authors' opinions to construct their own position, for instance, by counter-claiming or pointing out gaps in existing research (including, of course, the book under review).Thus, direct criticism seems to be clearly expected in the genre.However, a significant difference exists in terms of frequency.A higher rate of direct criticism is observed in English.This contrasts with the low rate of direct criticism (35%) found in the Italian data set, and accordingly the more frequent use of mitigated criticism.That supports previous studies that Italian academics are far less likely to express criticism (Giannoni 2006).As Giannoni rightly observes, this may be interpreted as related to "smaller academic languages" (Čmejrková & Daneš 1997;Holliday 1999), that tend to reflect tight social networks and a less assertive stance, with an emphasis on collaboration rather than competition.On the other hand, the greater use of direct criticism in English academic prose may depend on the greater pressure -referred to as "fierce competitiveness" (Duszak 1997, 32) or "academic struggle" (Čmejrková & Daneš 1997, 52) -imposed upon the members of British academia, thus confirming previous studies (Salager-Meyer et al. 2003).
Another possible explanation may be related to different cultural writing conventions.
Previous studies have suggested that critical reviews are often encouraged in English linguistics journals (Itakara & Tsui 2011).This is also the case with my data sample, where the review editorial policy of the journals under scrutiny emphasises the importance of reviews as critical discussions.Unlike English, the Italian journals under investigation do not give any explicit instructions about reviewing.This lack may account for the more frequent use of mitigated criticisms in the Italian data set.Italian reviewers seem to feel more uncomfortable with direct criticism of other academics' work as, without specific instructions, direct criticism may constitute a face-threatening act.
These differences in reviewing practices between English and Italian confirm a previous study (Diani 2015), in which it was found that Italian historian reviewers tend to avoid criticism and that they prefer mitigated criticism to direct criticism in reviews.The fact that Italian linguistics and historian reviewers display similar degrees in the evaluation of scholarly writing proves that academic writing is "a collective social practice" (Hyland 2000, 1), which reflects the values, norms, conventions, beliefs or attitudes shared by a community.
Some implications can be drawn from this study for the field of EAP pedagogy.As research has shown (Kiely 2004;Giannoni 2006), understanding evaluative speech acts is necessary for pedagogic purposes, considering that critical speech acts are important in disciplinary gatekeeping.Providing learners and practitioners with the lexical and pragmatic dimension of evaluative speech acts in book review genres may develop their language awareness and encourage them to produce review texts and thus become active members of their discourse community.

( 9 )
This first problem may be confusing for some inexperienced readers.(EN_LIBRA) (10) This point might at first seem hardly worth mentioning.(EN_LIBRA) 5

) <…> But Lightfoot has two real problems here. The
contains some of the traditional sections of a research article: Introduction, Discussion, Conclusion.The Introduction and the Conclusion are often very similar in structure to those of the research article, as shown in ( Römer's (2005)or each sub-corpus was generated, from which inherently critical expressions carrying negative evaluation were identified, e.g.Engl.failure,objection;Ital.obiezione 'objection', mancanza 'lack'.By carrying out this initial analytical step, different types of lexically based criticism were found and grouped according to word classes, usingRömer's (2005)criteria adopted in her study on the expression of negative evaluation in book reviews.

Table 1 .
Type and distribution of mitigation strategies for criticism in English linguistics book review articles, from the EN_LIBRA sub-corpus 3

Table 3 .
Type and distribution of mitigation strategies for criticism in Italian linguistics book review articles, from the IT_LIBRA sub-corpus

Mi dispiace che l'analisi dei dati non sia corredata da un glossario, che avrebbe sicuramente giovato a
una più esatta interpretazione dei dati.(IT_LIBRA)'Iregret that the data analysis lacks of a glossary which would have benefited from a more accurate interpretation of the data.'As the above examples indicate, criticisms combine with suggestions, realised through a hypothetical clause including the conditional sarebbe 'would be' or avrebbe 'would have'.Once again, the full force of the criticism, focusing on the weaknesses of the reviewed author's book (l'autrice usa un ristretto numero di studi scientifici 'the author uses a limited number of studies'; mi dispiace che l'analisi dei dati non sia corredata da un glossario 'I regret that the data analysis lacks of a glossary'), is softened by a suggestion offering directions (sarebbe stato più produttivo allargare 'would have been useful to enlarge'; un glossario che avrebbe sicuramente giovato a una più esatta interpretazione dei dati 'a glossary which would have benefited from a more accurate interpretation of the data').Though limited in number, like the English sub-corpus, suggestions not only occurred in a 'criticism-suggestion' pair.They also occurred in statements of praise, as shown in (19) and (20).Entrambe le analisi lessicali sono corredate da tavole sinottiche e belle presentazioni cartografiche.Per il futuro sarebbe auspicabile confrontare il quadro tracciato con la realtà linguistica di altre comunità vallesi.(IT_LIBRA)'Both lexical analyses contain synoptic charts and nice maps.Henceforth, it would be worthwhile comparing these results with the linguistic situation of other Valais communities.'Anotable difference between the two corpora was found in the frequency of hedges used as mitigation strategies.Italian texts contain only 4 occurrences of hedges.In terms of lexical choice, however, the results show a remarkable similarity.Like the English subcorpus, the hedges used in the Italian texts are the modal lexical verb sembrare 'seem' (3 occurrences) and the adverb forse 'perhaps' (1 occurrence): (21) L'affermazione dell'autrice sembra piuttosto vaga <…>.(IT_LIBRA) 'The author's statement seems rather vague <...>' (22) Forse non è del tutto corretto, come fa Eco, affermare che <…>.(IT_LIBRA) 'Perhaps Eco is not quite correct when he claims that <...>'