Keywords

1 Introduction

Driven by events of international scope hosted in the countryFootnote 1, as well as worldwide recognition of Brazilian chefsFootnote 2 who favor local ingredients in their recipes, Brazilian cooking has attracted growing attention, and as a result a number of cookbooks have been either translated into English or written in this language.Footnote 3 Nevertheless, the theme lacks reliable reference material that could help translators and those who write in this specialized area. If indeed there are few studies devoted to culinary writing as a whole [2], Brazilian cooking suffers from the same neglect.

Strong cultural aspects such as local ingredients, techniques and utensils of Brazilian cuisine are usually reduced to a small number of entries in general language dictionaries or are restricted to lists of terms aligned to possible equivalents. Not only are these sources insufficient to account for much of the Brazilian cuisine terminology, but they also err in not contextualizing these terms, which can mislead translators and writers on how to use them. Because cooking is often considered as not belonging to a specialized area, and the translation of culinary texts may be delegated to either nonprofessional translators or automatic tools, the consequences range from lack of standardization of equivalents - especially of cultural references - and naturalness, to plain errors, which obviously hinder reader’s understanding [18]. Hardly would a foreigner understand that a dish called ‘boyfriend in the oven’ refers to a type of roasted fish, or that ‘bread with cold’ is a cheese and ham sandwich, to number but a few of recurrent translation inadequacies found in bilingual restaurant menus and displays used to identify dishes served at self-service restaurantsFootnote 4.

However, as previously mentioned, equivalence of terms is far from being the only obstacle to the proper task of translating and writing about food. The way these terms are used in the texts highlights other gender specificities in different cultures. Finding in a list of ingredients expressions such as cabos de salsão [sticks of celery] and cebola cortada em quartos [onion cut in quarters], instead of the conventionalized talos de salsão [celery stalks] and cebola cortada em quatro [onion, quartered], respectively, would surely intrigue a reader who has modest knowledge of cooking.

This study aims to identify recurring phraseologies in authentic recipes written in Portuguese and in English in order to propose functional equivalents which could assist translators and those who write about cooking to produce more natural texts. The final product is a Portuguese-English dictionary of Brazilian cooking under construction, which is detailed in [18].

2 Translation, Terminology and Corpus Linguistics

Although autonomous, each with its own specific principles and purposes, specialized translation and terminology converge in many ways. Just as specialists rely on glossaries to ensure accurate communication in their fields, translators use this lexicon to convey specialized knowledge from one language/culture to another. Nevertheless, despite targeting unambiguity and homogeneity, specialized language is far from unequivocal, and equivalence of specialized lexicon is not the only aspect to be considered in translation. Contrary to traditional studies in translation, the specialized text is also subject to the cultural context of production and reception, and these elements should also be observed during the translation process [23].

When differentiating between ‘intention’ and ‘function’ as discussed in translation functionalist approaches, Nord [15] points out that while intention is defined from the sender’s point of view - good intentions do not guarantee an adequate translation, though, it is the receiver who uses the text with a certain function “depending on their own expectations, needs, previous knowledge and situational conditions” [p. 28]. Therefore, a recipe, for example, must be seen by its user as such, which means that it needs to have the same characteristics that this genre in this user’s cultural setting has. In addition to that, equivalent terms do not guarantee an appropriate translation either, since words tend to appear in groups [8].

Specialized phraseology, or relatively fixed structures, has been receiving special attention in terminological studies (see, for example, [12]), making us aware of the difficulties of delimiting terms and phraseologies of specialized texts in a single language. Adapting them to the target language/culture is even more challenging, and that is why the translation of terms and phraseologies from one language to another needs to be adequate, aiming at conceptual precision through lexical views toward standardization [11], albeit in an idealized way, since there is no way to guarantee it in practice.

Just like any other specialized text, a recipe contains lexical and syntactic specificities [3]; for example, the use of specialized vocabulary such as ‘cup’, ‘spoon’, ‘add’; and characteristic combinations of words, such as ‘finely chopped’, ‘stirring constantly’, ‘bring to a boil’, are immediately associated with the genre. In addition, the instructions, usually presented by verbs in the imperative such as ‘stir’, ‘let’, ‘mix’, at least in Portuguese and English languages, confirm the operational nature of the recipe [16]. Considering that specialized texts are repositories of terminology, we used the assumptions underlying Corpus Linguistics to assist us in the semiautomatic survey of phraseologies which are characteristic of Brazilian cooking and their respective equivalents in English.

The use of parallel corpora in translation is not a new practice. The comparison of original texts and their translations enables the identification of previously used translation equivalents in a relatively simple way, through the alignment of the sentences of the original text with those of the translated text. However, translated texts do not always reveal terms and phraseologies used conventionally in the target language, since they have been mediated by a translator, who may not be an expert in the area for which s/he translates. Comparable corpora, on the other hand, can reveal terms and phraseologies used naturally in the source language, and also help in the identification of discrepancies between textual types produced in different languages and cultures [17]. As specialized texts, recipes can be used for the extraction of terms and phraseologies, thus contributing to the construction of bi- or multilingual terminology reference material. Next, we describe how the authentic texts served the purpose of this research.

3 Study Corpus: Compilation and Analysis

The criteria underlying the choice of the study corpus should be guided by the objectives to be achieved. Thus, we tried to be very specific in the choice of the texts representing both languages in order to meet our objective of identifying typical phraseologies in Brazilian recipes and their functional equivalents in English. The corpus was compiled with recipes from twenty two Brazilian cookbooks, eleven of which were originally written in Brazilian Portuguese and eleven in North-American EnglishFootnote 5. Table 1 gives a summary of the study corpus:

Table 1. Study corpus
Full size table

There is no consensus regarding the ideal size of a study corpus [22]. Sinclair [20] argues that a corpus should be as extensive as possible; whereas Koester [10] claims that more important than its size is its representativeness. Although not very extensive, specialized corpora, which are commonly used for conducting specific research, may lead to quite significant results, as long as they have been built under strict criteria, always considering the purpose of the investigation.

Koester [10] points out that small corpora have the advantage of allowing a closer relationship between the corpus and the contexts in which their texts were produced. In addition, as in research with small corpora the compiler is usually also the analyst, the degree of familiarity with the context tends to be higher, allowing the quantitative data revealed by the corpus to be complemented by manual analysis. Although it is difficult to guarantee that a particular corpus is ideal for any research (See [13]), and since our study corpus is very specific, we believe that its relatively small size was not an impediment for the analyses carried out.

In this research we are concerned especially with the recurrent phraseological units in Brazilian recipes. So, using WordSmith Tools 6.0 [19] we identified the most significant combinations, i.e. key-clusters, in the study corpus by comparing Brazilian recipes to general recipes, the latter accounting for over 1,000,000 tokens, being used as our reference corpus. The processing of key-clusters is similar to the one of keywords, in such a way that the tool presents us with the most statistically recurrent combinations in the study corpora when compared to the reference corpora. Tables 2 and 3 respectively show the first twenty key-clusters identified in Portuguese and in English in decreasing order of keyness, considering combinations of three to six words which recur at least ten times in the corpora.

Table 2. First twenty key-clusters in Portuguese.
Full size table
Table 3. First twenty key-clusters in English.
Full size table

It seems important to emphasize that the tool solely shows statistical results based on recurrence. It is the researcher’s task to distinguish between combinations of words that do not constitute complete units of meaning, e.g. de farinha de mandioca [of manioc flour], sem pele nem [peeled and], ‘to a boil’, ‘cook over low’, etc., from the ones that form complete units of meaning, be they compound terms, collocations or phraseologies such as [xícara de chá de] farinha de mandioca [cup manioc flour], [tomate] sem pele nem sementes [tomato, peeled and seeded], etc.

Naturally, had the corpora been POS-tagged, the results could have been optimized, since the processing could start from content words, ignoring grammar words, for instance. Nevertheless, due to the size of the corpora, it was possible to manually analyze the clusters presented by the tool, along with the analysis of those combinations in the context they appear by using the ‘concordance’ tool, in order to identify phraseological differences in the two languages. This will be explained in the next section.

4 Brazilian Recipes in English and in Portuguese: Phraseological Differences

A recipe is usually “incomplete, inexact and inconsistent” [16, p. 174]. For example, in instructions such as ‘stirring constantly’ and ‘egg whites, beaten’, it is assumed that, respectively, a spoon and a mixer, or other similar utensils, will be used so that the tasks are accomplished. Diemer [6] explains that, historically, recipes were shared among cooks and, therefore, dispensed details. Over time, this textual genre expanded its scope to include the layperson and, as a result, the lexical complexity was gradually minimized, and more precise measures and more detailed information were introduced. However, incompleteness still underlies this textual genre, albeit in varying degrees, in different languages/cultures. When comparing recipes in four variants - American and British English, and European and Brazilian Portuguese - Tagnin and Teixeira [21] concluded that Portuguese and British recipes tend to be more precise in terms of measurements, using ounces and pounds, for example, while those from the ‘younger’ countries - Brazil and the United States – prefer the ‘looser’ spoon(s) and cup(s). Such differences have an impact on the translation and writing of texts in the area, at least when naturalness and conventionality are sought in the translated text. Next, we will detail differences found in the study corpora and their consequences for translation.

4.1 Inaccuracy

While reporting his insecurities as a casual cook, Barnes [1] complains that cookbooks are not accurate. Referring to a British cookbook, he mentions the problem with onion size, which, in these books appears to have only three - small, medium and large, when in reality it may vary quite substantially. Had this British cook/writer analyzed Brazilian cookbooks written in Portuguese, he would certainly have been even more perplexed by the level of imprecision found in this material.

Among the keywords of the subcorpus in Portuguese, we were struck by the adjective bom/boa [good]. We then searched for their contexts, using the Concordance tool. Figure 1 shows a sample of this search.

Fig. 1.
figure 1

Sample of the concordance lines with the search word bom in the Portuguese subcorpus

Full size image

By analyzing the combinations with bom/boa, we concluded that, besides referring to quality, this adjective is used to refer to what is ideally expected in terms of size, consistency and quantity in the recipes in Portuguese. For example, Fig. 1 shows occurrences such as bom espaço [good space], bom pedaço [good piece], bom molho [good sauce], and bom tamanho [good size]. In the English subcorpus, on the other hand, similar concepts are expressed by adjectives such as ‘large’ and ‘thick’, e.g. ‘large green bell pepper’ and ‘make a thick sauce’, whereas ‘good’ is used mostly to indicate high quality as in ‘a good substitute’ and ‘a good chocolate’ (see Fig. 2):

Fig. 2.
figure 2

Sample of the concordance lines with the search word ‘good’ in the English subcorpus

Full size image

4.2 Implicit and Explicit Information

The level of explanation of the instructions to be followed in recipes also differs in recipes written in Portuguese and in English. For example, among the key phraseologies in Portuguese we have limpe os camarões [clean shrimp], in which the steps involved in the process remain implicit. In English, a recurring phraseology is ‘shell, devein and wash/clean shrimp’.

Another example of explicitness refers to caution in the English subcorpus: ‘using a pair of tongs, remove * from the pan’. In the subcorpus in Portuguese, instructions are restricted to the action – retire * da panela [remove * from the pan], not mentioning how the action should be carried out. Thus, we can conclude that in addition to the more detailed description of the process, there is more concern with safety in the English texts, as the cook is instructed to use appropriate utensils and to be careful when handling hot ingredients.

4.3 Technicality

Although recipes can be classified as technical texts, gastronomic material is aimed at different users, such as laypeople, chefs, children, newlyweds, among other enthusiasts; additionally, recipes are also shared in blogs, with their own specificities (See [7]). All these variables certainly interfere in the choice of the terminology used in the texts. In his linguistic analysis of recipes, Cotter [5] concludes that those aimed at laypeople are more explicit and detailed than the ones aimed at professional chefs, for example. But in addition to the aforementioned variables, the degree of technicality in the recipes also varies according to the language/culture in which they are produced.

Having compiled a comparable corpus which is homogeneous in what regards (i) mode – written cookbooks, (ii) authorship – native speakers of the languages involved, (iii) medium – printed, (iv) target audience – laypeople, and (v) amount of texts – eleven cookbooks in each language; we also expected to identify a balance in relation to the degree of technicality used. However, results showed that it was not always the case. For example, significantly recurring in the English subcorpus, the term ‘mandoline’ is used in phraseologies such as ‘cut * (into julienne), using a mandoline’ and ‘thinly slice * (preferably) using a mandoline’. In the Portuguese subcorpus, the prima facie equivalent – mandoline(a) – has no hit. In order to designate a similar concept, the Portuguese subcorpus offers cortador de legumes (vegetable cutter), yet with only two occurrences. In general, recipes in Portuguese do not specify the utensil with which the task should be carried out.

Another term present only in the English subcorpus is ‘cheesecloth’, found in phraseologies such as ‘allow to cool in the cheesecloth’ and ‘place the mixture in cheesecloth and squeeze’. Similar instructions are given in Portuguese by using pano (bem) fino [(very) thin cloth] – ponha * sobre um pano bem fino [place * on a very thin cloth], esprema bem em um pano fino [squeeze well using a thin cloth]. A more technical term – morim – was not found in the Portuguese subcorpus.

Another commonly mentioned utensil in the subcorpus in English is cocktail shaker, part of the combinations ‘shake well in/transfer to/pour into the cocktail shaker’. Its prima facie equivalent in Portuguese – coqueteleira –, on the other hand, is not used at all in the Portuguese subcorpus. We can infer, therefore, that the use of more technical terms in cooking phraseologies in English is considered acceptable by North-American nonprofessional cooks, but not by their Brazilian counterparts, as they may interpret it as excessive use of technicality.

4.4 Semantic Differences

Another feature that differentiates phraseologies used in the recipes in Portuguese and English refers to the semantics underlying the verbs characteristic of this textual genre. In English, verbs used to describe cooking are usually imbued with full meaning, such as ‘dice’, ‘julienne’, ‘slice’, ‘half’, ‘quarter’, ‘peel’, used in phraseologies such as ‘tomatoes, peeled and diced’; ‘olives, pitted and sliced’; and ‘onion, julienned’. Similar instructions in Portuguese are frequently expressed by verb phrases with cortar [cut]: cebola cortada em rodelas [onion cut in slices]; batatas cortadas em cubos [potatoes cut in dices]; corte os palmitos ao meio [cut the palm hearts in half]. Even when there is a single verb in Portuguese to express a certain action, in recipes it seems that preference is still given to verb phrases. For example, the Brazilian Portuguese prima facie translation for ‘slice’ is fatiar. In the recipes, however, the verb is used basically as a participial adjective, in sentences such as bacon fatiado [sliced bacon] and maçã fatiada; the noun fatia is also used, as in fatia de presunto/queijo/pão [slice of ham/cheese/bread].

5 Different Cultures, Different Amount of Information

Drawing on the anthropologist Edward Hall’s concept of ‘contexting’, Katan [9] claims that “[…] individuals, groups, and cultures (and at different times) have differing priorities with regard to how much information (text) needs to be made explicit for communication to take place.” [p. 177]. The researcher designates ‘high context cultures’ those which rely more on prior knowledge, thus requiring less text to convey information. Low context cultures, on the other hand, assume that common knowledge is insufficient to guarantee communication, demanding more explicitness. Figure 3 summarizes the contexting cline:

Fig. 3.
figure 3

[Adapted from Katan, 1999]

Context classification of cultures

Full size image

According to the cline, Latin American cultures, including Brazilian culture, are higher than North-American in relation to contexting, requiring fewer words to accomplish successful communication. This helps explain why the instructions found in Brazilian recipes written in English, as shown above, are more detailed than those written in Portuguese, corroborating Navarro’s [14] assertions. Using Corpus Linguistics to analyze hotel sites in Brazil and the United States, the researcher concluded that the latter are more detailed. In relation to recipes, not surprisingly, we reached similar results since our English subcorpus is composed of books written by North-Americans. We did not research other English speaking cultures to verify whether they fit into either low or high context cultures because it would escape the scope of this study

Naturally, other factors may influence the amount of information revealed or implied during communication. Recipes aimed at a foreign culture are expected to bring more comments on the local ingredients and preparation, for example. Even the choice of recipes may be different (See [18]). However, since the two subcorpora compiled for this study focus on nonprofessional cooks, we believe that the difference between the amount of information expressed in the phraseologies and the degree of technicality of terms may relate to the different languages/cultures involved.

In order to assist translators and writers in producing recipes that sound natural to English-speaking readers, the phraseological differences herein discussed is taken in account in the creation of a bilingual terminological material in this specialty area. In order to illustrate the proposed translator-targetted dictionary of Brazilian cooking, we present below the entry camarão [shrimp]. In addition to the fields dedicated to translation equivalent (TE.), definition (Def.), example (Ex.), and comment (Comment.), we included phraseology (Phrase.), aiming at facilitating the user’s choice of a functional equivalent in English (Fig. 4).

Fig. 4.
figure 4

Dictionary entry camarão

Full size image

6 Concluding Remarks

Although interest in cooking has been on the increase all over the world, reliable reference works in the area are still scarce. As far as Brazilian cooking is concerned, reference materials in the Portuguese-English language pair are non-existent. Consequently, texts of this domain written in English or translated into this language show a series of problems, among them multiplicity of translation equivalents, lack of characterization of cultural references and lack of naturalness in the target language.

Due to lexical and syntactic specificities, a recipe is a repository of terminology and phraseology typical of the area and, therefore, constitutes appropriate source for the construction of a bilingual terminological resource for translation purposes. The semiautomatic analysis of the corpus of study, namely Brazilian cookbooks written in Portuguese and in English, revealed, in addition to simple and compound terms, prototypical phraseologies of the recipes produced in both languages. This survey, which combines qualitative and quantitative analyses, evidenced differences that go beyond terminological equivalence.

Although incomplete, since they presuppose prior knowledge on the part of the reader for its correct accomplishment, this incompleteness is unequally revealed in recipes in the two languages/cultures herein analyzed. The processes underlying Brazilian recipes in English indicated a greater tendency for detail and warning than the Portuguese processes, corroborating Katan’s [9] conclusion that North-Americans need more text in order to achieve understanding. In contrast, verbs in English used in recipes are more semantically charged than those used in this genre in Portuguese, a language which makes higher use of verbal phrases in recipes, even though it has single verbs that play the same role. In addition, the degree of technicality also varies in this textual genre intended for similar audiences. The aforementioned aspects corroborate Zavaglia et al.’s conclusions [23], according to which the technical text goes far beyond terminological equivalence, since it involves other variables that must be adapted in the target language/culture so that the translation results in a fluent text.

In addition to the compilation of a Portuguese-English dictionary of Brazilian cooking, which we expect to directly assist translators and those who write on gastronomy, our study could also have an impact on the formation of specialized translators. Translation students must be made aware that, when translating a recipe into English, they have to consider that this language requires more detail [as highlighted in Katan’s cline], that special attention must be paid to variation in degree of technicality, and that cultural aspects are as important as linguistic ones. By so doing, we believe that they will be able to produce more adequate texts; texts that will be better accepted by target audiences.