Corps de l’article

1. Introduction

The lexicon is undoubtedly a prominent level of analysis in the search for patterns typical of translated language, or features of translation (term borrowed from Olohan 2004: 92). The focus has been placed mostly on measures of frequency and distribution of lexical items to test for hypotheses such as (lexical) simplification (Blum-Kulka and Levenston 1983; Laviosa 1998; Corpas Pastor 2008) and lexical normalization (Kenny 2001; Hansen-Schirra 2003). Despite the reality of translation as manipulation of meaning, studies using data from lexical meaning are less abundant. Within the realm of compositional lexical meaning, an under-explored area of interest to inform on features of translation is the opposition abstract vs. concrete.[1]

This paper aims to contribute to the existing knowledge about changes in abstractness in translated texts; specifically it proposes and tests a hypothesis for the Spanish into English language direction, namely that abstractness is reduced because fewer abstract words are used in the English versions. Abstractness, defined below, is operationalized as the occurrence of a class of abstract nouns, those denoting properties of entities, for example “elegance” in “the elegance of a tall hedge,” as in Example 1 below. Properties are types of meanings that are used for evaluation, specifically in the expression of the speaker’s attitude in the domains of ethics, aesthetics and emotion (see, for instance, Martin and White 2005: 40).

The chosen genre is fictional prose (Biber and Conrad 2009: 19), in particular two novels by Spanish writer Manuel Vázquez Montalbán (1939-2003), Los mares del sur[2]/Southern Seas[3] and Tatuaje[4]/Tattoo.[5] In fiction, evaluation contributes to delineating narrative point of view: “To tell a story from a character’s point of view means to present the events as they are perceived, felt, interpreted and evaluated by her at a particular moment” (Niederhoff 2011/2013: 5). Consider Example 1, a description of a mansion’s gardens where nouns and adjectives are used to evaluate the landscaping.[6]

The narrator is describing the gardens from the point of view of the character that is paying a visit, middle-aged detective Carvalho. The observer’s attention, which the reader may construe from lexical choice and syntactic articulation, is focused on the formality of the landscaping. The English translation is literal (Delisle, Lee-Jahnke, et al. 1999: 154), except for the replacement of a relative clause with an independent sentence. In the second sentence, the recurrent analytical grammatical pattern of a head noun with a prepositional phrase complement (“the elegance of a tall hedge”) is preserved.

The description proceeds with the driver in a subtle ironic tone. This is reproduced in Example 2, with property-denoting nouns in bold.

In the Spanish version, qualities of objects are presented through the abstract nouns contención [restraint] and (mucho) mundo [worldly wisdom]. In the second sentence, which translates literally as “he appreciated the stylish restraint of his grey suit,” the property contención becomes grammatically foregrounded as the head of the noun phrase while the physical object of which the restraint is predicated surfaces in the complement. Hence, the character is reported to appreciate the restraint in the uniform, rather than the uniform.

In the English version of Example 2, information is packaged differently. Both abstract nouns are deleted; the verb appreciate has concrete objects as arguments (uniform and hands) and the entire sentence foregrounds the sensory aspects (fine, whitish-grey, velvet, leather, etc.). The adverb elegantly fails to express the cognitive assessment contained in “worldly wisdom,” a quality that presumably only a seasoned detective could project on the chauffeur just from the look of his hands. Such an insight would help the reader construe the scene from Carvalho’s point of view.

As Examples 1 and 2 illustrate, the author exploits the structural possibilities and stylistic preferences of Spanish to create a link between property attribution, evaluation and point of view. Instances of abstract nouns for property attribution abound throughout the novel; this can be interpreted as a motivated choice with literary relevance (Halliday 1971: 339). However, as the translation of Example 2 suggests, their frequency in English may decrease. As with any potential property of the translated text, the causes must be sought in a combination of factors. Steiner (2001: 5) proposed “typological factors, register, and properties of the translation process, understanding in particular.”

As regards typological factors, we can first draw on the insights about the expressive preferences of languages which were applied by work in the comparative stylistics discipline. In Vinay and Darbelnet (1995: 51), “abstract words often refer to generic concepts and tend towards the characteristics of mathematical symbols and consequently appeal to the intellect rather than the senses. Concrete words refer to physical objects or actions.” English is claimed to have a preference for concrete expression, that is, “a level at which the linguistic realisation mirrors concrete reality,” while French prefers abstract expression, a level which “shows reality in a more general light” (Vinay and Darbelnet 1995: 51). Vázquez Ayora (1977: 82-84) offers numerous examples of shifts from concrete in English to abstract in Spanish, in support of the resemblance of Spanish to French.

Case studies of translations support this view. Munday (2008: 129-120) reports a preference for “concrete translations” and a “concrete language” in the case of Cien años de soledad[7] by Nobel Prize Laureate García Márquez and its English translation[8] by Gregory Rabassa. Gesner (2004) observes semantic shifts from the abstract and general to the concrete and particular in French-to-English literary translation, although the opposite kinds of shifts are also observed, even if to a much lesser extent.

The alleged preference for concrete modes of expression in English seems to be reinforced by the linguistic styles prevalent in present-day fiction regarding the use of a specific grammatical category, namely nouns. Francis and Kučera (1982: 546) and Biber, Johansson, et al. (1999: 65) observed a tendency to use fewer nouns than verbs, at least in their fiction subcorpora, compared to non-fiction. Furthermore, the patterns of noun phrase postmodification have been evolving since the 18th century both in fiction and non-fiction; relative clauses and of-prepositional phrases have decreased in use, with a “steadily decreasing use” of the latter in fiction (Biber and Conrad 2009: 169). Thus, the rendering “the stylish grey uniform” matches the evolution of style preferences for fiction in English, which integrate the expectancy norms under which translators work (Chesterman 1997/2016: 62).[9] However, part of the meaning is lost.

The status of translation in the English literary canon works against the preservation of the abstract nouns in Example 2. Generally, translated literature in English is subject to a norm of fluency (Venuti 1995), regardless of a particular work’s position as high literature or popular fiction. The translation excerpted in Example 2 appeared in a crime fiction series from a small British imprint (Serpent’s Tail), although the source text had been published in a general fiction collection.[10] Translations of crime novels often show a lack of awareness of the genre-specific challenges, which include “formal and rhetoric peculiarities” that contribute to reader involvement (Hopkinson and Seago 2018: 227), namely the lexicon of emotion, traces of language varieties, and certain narrative modes like free indirect speech (see, for instance, Cadera and Pavić Pintarić 2014). Carter (2014) found fewer metaphorical expressions in English translations of French detective fiction compared to their source texts.

The third factor, the process of translation, “has to be seen in the context of and in interaction with the other two” (Steiner 2001: 5). Hansen-Schirra (2003: 297), comparing fiction portions of the Translational English Corpus[11] with the British National Corpus,[12] found that typical fiction features (such as time and place adverbs, perfect and past tense forms) are overused in translations, rendering translated fiction “more narrative and situation-dependent.” Interestingly, the study also found an overuse of untypical fiction features (nominalizations among them), suggesting that translations “point towards a more neutral register.” The role of genre has been observed in other studies of patterns of nouns and noun phrases in translational English. Alves, Pagano, et al. (2010: 116) report the percentages of category shifts in various combinations (verb-noun, noun-verb, adjective-noun, noun-adjective, etc.) comparing the directions English-German with German-English. Shifts from any category to noun occur in lower percentages in translated English than in translated German in register-controlled corpora of corporate communication. These figures suggest that “English translations are less nominal than their German originals” (Alves, Pagano, et al. 2010: 116). Lewandowski (2015), in a study on translations from Polish into English, found a decrease in noun phrase constituents together with a “concentration” strategy, which the author links to a general tendency observed by Delisle, Lee-Jahnke, et al. (1999: 127) in English translations from many languages. On the topic of abstract words, Obeidat (1998), in a small case-study of Arabic-English translation of literary texts, found that the percentage of abstract nouns went from 31% of all nouns in the Arabic original to 24% in the English translation.

These observations justify further research with the general aim of confirming reduced abstractness in English translations of Spanish fiction. They also stress the need for a definition of abstractness that teases out the semantic dimension (concrete vs. abstract) from the grammatical (its association with the use of nouns), in order to ascertain the factors leading to such a reduction and their potential link to different translation tendencies.

The paper is structured as follows. Section 2 is devoted to preliminary definitions. Section 3 presents the methodology of the study, namely data selection and classification of translation procedures to take abstractness into account. Section 4 provides a quantitative description of the translation correspondences in both works followed by a qualitative analysis. Section 5 provides a general discussion of the results, the potential effects of the shifts in abstractness for the narrative value of the novels and suggestions for further work.

2. Preliminary definitions

2.1. Abstract nouns and properties

The terms abstract and abstractness may designate different objects of study, and the defining criteria vary according to disciplines and approaches within them. This means that there are “ways to be abstract,” in the words of Zamparelli (2020: 202) in his review of the concept from a predicate semantics perspective. For the goals of the present study on the potential translation shifts undergone by nouns that attribute properties, two uses are relevant, namely conceptual abstractness (vs. concreteness), associated with the referential meaning of words, and grammatical abstractness, related to the choice of the grammatical category noun to express properties, by comparison with other categories, like adjective and verb.

For a definition of conceptual abstractness, I turn to the discipline of cognitive psychology, where abstractness of concepts is a valid operative parameter for a number of task-based experiments (Wiemer-Hastings and Xu 2005). Concepts are ranked on a scale from highly concrete to highly abstract (see, for example, Coltheart 1981). Highly concrete items are characterised by their physicality and by intrinsic and perceptual features; in contrast, abstractness correlates with a proportion of introspective features linked to mental experiences (emotion, representation, evaluation) and social properties (social state, social artefact, action, event). In Example 2 above, the words uniform, elegance, and worldly wisdom illustrate three points in the concrete-abstract continuum. The first item denotes an object and hence is the most concrete; the second item denotes a quality perceived in the colour contrast between the gloves and the wheel of a car; it is abstract though still based on perceptible properties. Finally, the last item denotes a cognitive state imagined by the narrator; hence it is the most abstract. Since the lexical choices of translators may introduce variations on this scale, I will consider conceptual abstractness a valid tertium comparationis for the classification of translation solutions.

The second definition of abstractness is related to the grammatical semantics of nouns. Adjectives “typically denote properties” (Huddleston and Pullum 2002: 527), such as wise, in “a wise teacher”); but properties can be expressed by words belonging to other grammatical classes, such as nouns (for example, “the teacher’s wisdom”). For a specific subset of nouns, namely those that have formal adjective counterparts, some scholars propose making finer distinctions by taking into account their embedding structure, that is, the noun phrase.

While the noun wisdom denotes a property, “the teacher’s wisdom” denotes a specific manifestation, an instance of property attribution (a “trope,” in Moltmann 2004). One of the reasons for drawing such distinctions is the observation that nouns derived from adjectives can participate in further predication relations (for instance, “students respect the teacher’s wisdom”), which makes them intuitively “object-like.” The nominalisation process is considered a mechanism to reify the contents of a predicate, in other words to allow reference to them as an object (Iturrioz Leza 2000: 89). Furthermore, the fact that reified properties occur as heads of noun phrases (and hence may take on argument functions and semantic roles) allows them to build more abstract descriptions and representations of reality in texts than their counterparts from other categories.

In systemic functional linguistics the reifying potential of nominalisations derives from the phenomenon known as “grammatical metaphor,” whereby the encoding of a process as a noun is a more indirect way of mapping semantic roles onto grammatical functions than when processes are encoded as verbs (Halliday and Matthiessen 1999). For Steiner (2004: 138), the mapping of a process as a verb is more explicit, and demetaphorization, or rephrasing a nominalization in verbal form, “seems to model aspects of ‘ambiguity,’ ‘explicitation of meaning,’ and to that extent, ‘understanding’.” In a similar vein, Martin and White (2005: 150) propose that encoding qualities as nouns instead of adjectives, a different kind of metaphorization, adds semantic complexity.

For the purposes of the present study I have selected instances of nouns attributing a property to an individual or entity that is contextually anchored in the story time (Genette 1972/1980), more specifically, nouns ending in -idad, a variant of the derivational suffix -dad building “nouns of quality, state and condition” (Real Academia Española 2009: 413, my translation).[13] “Quality-denoting” nouns may refer both to intrinsic properties that characterise an entity (such as fidelidad [faithfulness], marginalidad [marginality]) and to transient properties (like agresividad [aggressiveness]) (Real Academia Española 2009: 340).[14]

2.2 Translation tendencies

According to the definitions of abstractness provided above, renderings of nouns in -idad (and the larger units that embed them) may vary in either conceptual abstractness or grammatical abstractness, or in both. Each level may in principle be affected by different translation tendencies.

One of the tendencies observed in translated English is simplification (Baker 1996). Laviosa (1998) studied lexical simplification in the Translational English Corpus by identifying, among other parameters, list head items. She found that “[o]n average, the proportion of high-frequency words used is significantly higher in the translated narrative than in the comparable original works” (Laviosa 1998: 6). The most comprehensive study of translational English using the methodology of linguistic variation analysis has confirmed that overrepresentation of the most frequent words and reduced information load (lexical density) are functional characteristics of translational English (Xianyao, Xiao, et al. 2019). However, frequency is only at best an indirect measure of simplification of conceptual abstractness, on the assumption that concrete words are more frequent than abstract words. More delicate distinctions are necessary, even if they do not lend themselves easily to automated analysis.

As regards grammatical abstractness, it can be argued that the rendering of property-denoting nouns is subject to a process similar to grammatical demetaphorization and the subsequent re-metaphorization in the target language (Steiner 2001: 15). This process is considered evidence for the explicitation/implicitation hypothesis, as not all grammatical metaphors are eventually reconstituted; some processes are rendered through more explicit verbal forms.[15] This was attested in a study of Portuguese to English translations performed by academic researchers (Silva 2007; Pagano and Silva 2008) and also in an experimental study involving professional translators vs. academics from Portuguese and German into English (Alves, Pagano, et al. 2014). As is the case with grammatical metaphor, the “un-reification” of properties should leave traces in translations in the form of more explicit renderings, namely as adjectives, verbs, and adverbs.

An overall factor to be considered is interference, stated as a probabilistic law in Toury (1995: 274-279), in which source language features are more likely to appear in the target text if the target language has them ( positive interference or transference) and features absent in the source language being less likely to appear in the target text (negative interference). Many -idad nouns have fully equivalent English cognates ending in the suffix -ity, a derivational suffix used to form abstract nouns meaning “state or quality of being A[djective]” (Biber, Johansson, et al. 1999: 322). Hence the existence of viable cognates may trigger a higher than normal frequency of such forms, an anti-normalization trend (Hansen-Schirra 2003) that would counter the loss of abstraction. On the other hand, words of Latinate etymology are considered more learned in English than in Spanish. For nouns ending in -ity, Biber, Johansson, et al. (1999: 322-323) report distribution patterns of 1,000 words per million in fiction and 5,000 words per million in academic registers. The law of growing standardization, defined in Toury (1995: 267-274) as the modification of textual patterns in the source text in favor of more common linguistic options by a target repertoire, may counter transference of -ity forms, a symptom of genre-related normalization.

These various tendencies may shape translation products differently as they interact with external factors such as translator style and sociocultural trends. However, if reduced abstractness is a robust hypothesis, it should be observable in both works analysed.

3. Method

The data for the present study is a set of Spanish-English concordances obtained through the following process. First, a search for items ending in -idad (singular) in the digital versions of the Spanish novels was conducted. Los mares del Sur contains 167 instances and Tatuaje contains 80 instances. After manual textual analysis of each instance, only those denoting properties of entities were selected (46 in Los mares del Sur, and 39 in Tatuaje).[16] Their English correspondences were identified, analysed in their broader co-text and classified according to the kinds of textual correspondences given to abstractness on both levels, conceptual and grammatical.

For conceptual abstractness, three categories are distinguished, namely those renderings that preserve the same degree of abstractness (including word-for-word translation and semantic shifts), those that reduce abstractness (including semantic shifts, insertion of support concrete nouns, and deletion of the segment), and those that increase abstractness, theoretically possible but not detected in the present study. This typology is presented and examplified in Table 1.

Table 1

Types of translation correspondences according to conceptual abstractness[17]

Types of translation correspondences according to conceptual abstractness17

-> Voir la liste des tableaux

As regards grammatical abstractness, a distinction is drawn between equally abstract textual solutions and less abstract ones, by comparison with the source text. I use the terms parallel, divergent and zero (Ø) correspondences, the latter two following Johansson (2007: 23-26). Hence, a correspondence is parallel when the property is rendered by an abstract noun; it is divergent when a class-shift (noun into adjective, adverb or verb) or a unit-shift (for example, noun into clause) – terms from Catford (1965: 80) – occurs. In a parallel correspondence, deletion of other items in the immediate co-text of the noun is not considered a reduction in grammatical abstractness. Table 2 summarises the typology of correspondences and provides examples.

Table 2

Types of translation correspondences according to grammatical abstractness

Types of translation correspondences according to grammatical abstractness

-> Voir la liste des tableaux

In a divergent correspondence, information is redistributed, often affecting the whole clause and even higher-ranking units. In the example provided in Table 2, the noun phrase headed by locuacidad (from the adjective locuaz, that is, being someone who talks in excess) is expanded into a clause with the verb talk as its predicate. The referential content – though not the speaker’s attitude – of the superlative premodifier nula is encoded by the modal of disposition want to and by clausal negation. While Spanish condenses the traveler’s subjective experiences (imposed silence and fatigue) and the resulting evaluation of the journey in one sentence, the English version uses three finite clauses and a factual sequential structure to narrate the travelers’ behavior.

The semantic function of the phrase can facilitate a divergent solution. For example, when the property noun is part of the manner adjunct <con [with] + noun>, a manner adverb might be a more direct encoding, even if a parallel template is available in English (such as con habilidad translated as skilfully instead of with skill).

4. Analysis of the translation correspondences

This section offers a description of the translation correspondences according to the typologies presented in the previous section (Tables 1 and 2). A quantitative description is provided first to assess the reduction in abstractness. A qualitative analysis of the typologies follows, with some remarks about the potential impact of shifts on narrative point of view.

4.1. Quantitative description

This is an exploratory study of how property-denoting nouns ending in -idad are rendered in English. The initial hypothesis, as we anticipated in the introduction and in Section 2.1, is that abstractness decreases. The raw numbers as well as the percentages for each type of abstractness are provided in Tables 3 and 4.

Table 3

Conceptual abstractness in Southern Seas and Tattoo

Conceptual abstractness in Southern Seas and Tattoo

-> Voir la liste des tableaux

Conceptual abstractness is reduced in both translations compared to their source texts, from 15.2% of cases in Southern Seas to 17.9% of cases in Tattoo. Zero translation, an extreme form of simplification, is the most frequent operation, although in Tattoo other solutions carry similar weight.

Table 4

Grammatical abstractness in Southern Seas and Tattoo

Grammatical abstractness in Southern Seas and Tattoo

-> Voir la liste des tableaux

Reduction in grammatical abstractness (divergent correspondences plus zero) is noticeable in Southern Seas (32.6%) and high in Tattoo (46.2%). The extent and means vary; class and unit shifts leading to unreified properties are more frequent in Tattoo (38.5%) than in Southern Seas (19.6%); in contrast, zero translation is more frequent in the latter. Regarding divergent correspondences, the most frequent group is nouns rendered as adjectives (5 instances in Southern Seas, and 9 in Tattoo), followed by adverbs (3 in Southern Seas and 2 in Tattoo) and the least frequent is nouns rendered as clauses (none in Southern Seas, 2 in Tattoo).

As regards the weight of interference, the percentage of Spanish nouns with English cognates that are potentially viable translations (that is, items ending in -idad with English correspondences ending in -ity) is practically identical (Southern Seas, 69.5% vs Tattoo, 70%). However, the percentage of English -ity nouns is higher in Southern Seas (39.1%) than in Tattoo (30%), suggesting a higher effect of transference in the former.

In spite of the limited number of instances, which precludes tests of statistical significance, the observation of reduced abstractness is encouraging. In the following subsection, qualitative text analysis helps elucidate the usefulness of the distinction between conceptual and grammatical abstractness for the design of a larger-scale study.

4.2. Qualitative analysis

In translation, shifts may affect one of the levels of abstractness or both. In this section, different combinations are analysed in detail.

4.2.1. Parallel conceptual and grammatical abstractness

Conceptual abstractness is preserved when an abstract noun is rendered as an abstract noun, as illustrated by the items in bold in Example 3.

Both versions establish a contrast between the feelings of security that the buses trigger in the onlooker (calor hogareño or homely warmth in Spanish, shelter in English) and the discomfort and insecurity produced by the morning cold, perceived as hostile. Conceptual abstractness is reinforced by grammatical configuration: the property hostility is foregrounded as phrasal head, while the physical attribute fría [cold]/sharp is a modifier and the notional argument is embedded as a complement. Hence, Example 3 is also an instance of parallel abstractness on the grammatical level.

4.2.2. Parallel conceptual abstractness, diminished grammatical abstractness

Conceptual abstractness can be preserved independently of the word class of the textual correspondence. Grammatical categories other than nouns, particularly adjectives, but also verbs and adverbs, adopt different semantic roles and syntactic functions that undo reification and hence encode less abstract relations between the participants. A typical example of this group of correspondences is Example 4.

In Spanish, Carvalho is involved in a continuous mental process, that of “contemplating” the dignity of a man apparently unable to afford fitting clothes. In English, he is involved in a punctual perceptual event: the noun sight denotes perception, and its complement pensioner seems the semantically congruent choice (people see physical entities rather than the qualities that they embody, such as dignity). On the narrative level, the point of view is still that of the detective; however, he comes across as a less reflective man, since intellectual activity is replaced by an involuntary percept. As the property surfaces as a premodifier (an adjective/participle), the recursive chain of complements (la dignidad [de un viejo jubilado [con chaqueta (…)]]) is lost; information about the jacket is provided in a non-restrictive relative clause, a syntactically free structure. Undoing reification simplifies syntax.

This reshuffling of semantic roles and syntactic relationships can facilitate access to concrete referents when lexical items are inserted. While in Example 4 the source text already provides a concrete argument for dignity (the pensioner), in several other instances translators have inserted a new item as the head for the adjective to modify. Some sequences are thoroughly rewritten, as in Example 5.

In this scene, the perceiver’s gaze focuses on the woman’s eyes, tearful but smiling. The manner adjunct is headed by the property noun serenidad [serenity], in turn complemented by the adjective troyana [characteristic of a Trojan]. The English sentence is rewritten more explicitly: a personal pronoun anaphor is chosen as the subject, and the manner adjunct is expanded with two new concrete noun heads, reducing grammatical abstractness. The noun look is the argument for stoic, and woman is the explicit argument of analogy-building Trojan. Furthermore, the impersonal situation of facing fatality is translated as accepting her destiny and death, which has the woman as the subject antecedent for the present participle. As regards the meaning of the property serenidad, the rendering stoic is not fully equivalent; nevertheless, it expresses a similar lack of display of pain as a judgement of the woman’s capacity, and hence retains the same degree of conceptual abstractness.[18]

The choice of a body part instead of the woman as the topic in Example 5 is a form of metonymy which lends an incongruous feel to the scene being depicted. The English version replaces the eyes with the personal pronoun she. Similarly, in Example 6, one clause is rendered by two finite clauses, each with a human referent subject.

The sentence is completely rewritten: the topic is no longer the figure at rest (Spanish formas en reposo) and the reference to her clothing is compacted in the noun air. The conceptual abstractness of the property dignity is preserved, but is predicated directly of the woman. Thus, the English version evaluates the woman in a simplified and more explicit way.

4.2.3. Diminished conceptual and grammatical abstractness

Diminished conceptual abstractness can result from a semantic shift, as lexical choice can fall on items lower on the abstractness scale argued for in Section 2.1. Only two such instances have been observed. Consider the example in Example 7, where the detective’s realisation of his own capacity to feel is replaced by the realisation that he is feeling something, an instance of affect.

The most extreme reduction in abstractness, both conceptual and grammatical, is the deletion of the segment embedding the property-denoting item, as in Example 8.

The deletion of phrases and even entire sentences might be a symptom of simplification under certain conditions linked to the genre of the work, particularly plot advancement. Once a novel is categorised as crime, action may be prioritized over scene description and social or cultural commentary. In Example 8, the trip to meet with one of the main witnesses provides the occasion to reinforce characterisation through a judgement of impropriety with mirón [ogler]. In Example 9, the detective shows a picture of the victim to the clients of a bar, young and old. The deleted sentence helps create an atmosphere of slowness by focusing metonymically on the tired look of the old clients. Such deletions may alter not only the construal of point of view but also the general mood by toning down intensity and removing descriptive details.

Deletion of lower-rank units (phrases) is found in both works. In Example 10, the translation replaces a relative clause with a postmodifier while removing one layer of description (the values of simplicity and taste). The passage is irrelevant for the plot; unfortunately, Vázquez Montalbán’s much celebrated knowledge of cuisine, reflected here in Carvalho’s judgement about a Catalan dish, is lost.

Deleted information is, arguably, recoverable from the situational context (Chesterman 1997/2016: 107) at least in some instances. In Tattoo, the phrase guiño de complicidad [a wink of solidarity] is translated as winking at him (verb). The gesture must have been judged as unambiguous in the situation as a thank-you gesture to the detective from a waitress in a bar.

4.2.4. Diminished conceptual abstractness, parallel grammatical abstractness

The insertion of concrete lexical items in the immediate context of the abstract word may contribute to reducing conceptual abstractness. When a property has some physical manifestation that allows the observer to detect it, the property-denoting noun occurs with a supporting noun (an air of, a display of), as illustrated in Example 11.

English and Spanish seem to differ in the usage patterns of such abstract nouns. I have considered such instances as conceptually less abstract, but grammatically equally abstract, as the noun is preserved.

5. Discussion and concluding remarks

The quantitative raw data show that both English translations resort to more concrete modes of expression than their source texts. Southern Seas does so to a lesser extent, as conceptual abstractness is reduced in 15.2% of correspondences, and grammatical abstractness is reduced in 32.6% of instances. Tattoo reduces conceptual abstractness in 17.9% of instances and grammatical abstractness in 46.2%. Despite individual differences these percentages lend support to the hypothesis of reduced abstraction in two works of translated crime fiction from Spanish into English. At the same time, the study does not find evidence for the semantic shifts from the abstract to the concrete claimed to be typical of translation into English by earlier case studies.

As can be seen in Tables 3 and 4, the types of textual correspondences vary.[19] The novel Southern Seas follows the source text’s grammatical patterns more closely, but resorts to deletion more often than Tattoo. This is somewhat contradictory, as the conservation of the source text’s abstract nouns at a microtextual level (that is, transference) is compensated by reducing text, normalizing the content of the novel in a way that is compatible with a given conception of the genre’s conventions. In turn, Tattoo resorts more often to divergent renderings with explicit participants and has fewer instances of zero translation. This could be attributed to both a stronger pull from the usage patterns of English fiction at the microtextual level (namely, standardization) and a tendency toward explicitation for the crime-novel reader.[20] From the perspective of style, transfer contributes to the consistent rendering of the source text’s textual patterns of evaluation, for example the combination capacidad de (capacity for) that construes the anticipation of a feeling or a judgement as the capacity to feel or to judge, as illustrated by the Spanish original in Example 8. There are 5 instances in Los mares del sur, 4 of which are translated literally, and one rendered through the French borrowing joie de vivre, of parallel conceptual abstractness. In Tatuaje, 4 instances were found. Two are rendered literally, one is rendered as the more conventional collocation “spirit (of sacrifice),” of equal abstractness, but different denotation, and finally Example 7 renders capacity as the actual occurrence of a feeling.

The group of property-denoting -idad nouns is heterogeneous. Properties range from “dispositional categories” associated with behavior (like locuacidad, talking in excess) to intrinsic states (like sensibilidad, the capacity to react emotionally). Divergent correspondences in general contribute to greater explicitness because as each property is unreified, its relationship with its argument is expressed through more “direct” phrase and clause structures (for example, behavior through verbs, and states through adjectives) and any implicit entities that make plausible arguments for those properties can be inserted.

The choice of property-denoting -idad nouns has proved very adequate to help establish a baseline for abstractness on different linguistic levels in Spanish-English translated fiction, even with a restricted set of data as the present one. More generally, it has contributed to a better understanding of the interaction between simplification and explicitation that had been established in the work on deverbal nominalizations/grammatical metaphor. The most dramatic rewriting takes place where reified properties cooccur with recursive complement structures and also where the semantic roles diverge from the concrete depiction of reality, as in Examples 4, 5, 6, 7 and 10.

Neither translation reflects the source text’s combination of abstractness and semantic incongruence, as both neutralize personification and metonymy, in line with the findings in Øverås (1998). The present study therefore supports earlier observations on the dilution of style in crime fiction translations. Two potential consequences for the reception of Vázquez Montalbán’s technique in particular may be anticipated. On the one hand, a reduction in abstractness enhances the “page-turner” quality of the book and its commercial advantage. On the other hand, since descriptive passages are retardation devices in the crime genre (Bayó Belenguer 2001: 91), translations adhere less to genre conventions. Perhaps a narrative advantage is gained with attention to physical detail, namely the representation of the narrator as objective and reliable rather than a figure that focalizes events through the mental activity of a character. As Venuti (1995: 287) states, “[i]f the illusion of transparency is strong enough, it may well produce a truth-effect, wherein the authorial voice becomes authoritative, heard as speaking what is true, right, obvious.” However, access to the detective’s state of mind and subjective representation of reality contributes to the richness of experience for the reader, a value that should not be lightly dismissed.