CHAPTER 13

There are three other categories which cannot be ignored in any discussion of Emotion, although they are not included in Ortony et al.'s scheme. One is the class of lexemes which refer to emotion generically, and other two are Surprise and Desire. The data on these groups is considered separately from that on the specific groups, so the results do not affect the general analysis of Emotion.

The generic lexemes of Emotion in English and Portuguese were described and compared in Chapter 2 with a view to explaining why emotion, or emoção, and not feeling or sentimento, has been chosen by psychologists to define this area of study. The use of these lexemes is not usually considered in normal lexical analyses of Emotion, but, since they belong to the metalanguage of Emotion, most authorities will attempt a definition of them at some stage.

Although these lexemes are generic when considered in isolation, in context they often combine with further specification of the emotion involved. For example, feel and sentir are complemented by adjectives, nouns and other parts of speech, as well as clauses, many of which specify the emotional content of the situation described. Sense of, feeling of, and sentimento de are also often complemented by specific Emotion nouns. These examples are considered here, because, although this complementation varies considerably, the generic word remains generic.

However, certain generic type lexemes can, in context, be recognized as belonging to more specific Emotion types despite the lack of complementation by an Emotion word. This is the case with temper, for example, which is usually qualified by bad or good, and therefore belongs to either the Anger or Joy groups. When the examples in question are truly specific, and occur in sufficiently high numbers, they have been classified under the respective group in the EC and PC.

However, the main problem is that so many examples from the main lexemes, feel and sentir-se combine with non-emotional meanings, or with expressions which have an underlying emotional connotation, as in:

where Sebastian uses this expression to convey an idea of physical and emotional depression. Feel also combines with a fair proportion of comparative constructions to describe situations that are emotional in content, as in:

(13.2) I felt like Lear on the heath, like the Duchess of Malfi bayed by madmen. BH

Rather than take too many individual decisions over which examples to include, I decided to exclude some less central examples- like the obviously physical examples of sensation and sensação - but did not cut any of the examples with the main lexemes feel and sentir. These lexemes were so central, and presented so many ambiguities, that it seemed a better idea to collect them all and then try and extricate the more salient facts.

Emotion itself is not a particularly popular word in either language, accounting for 6,6%(E) of the examples recorded for this group in the EC, with 5,9%(P) for emoção in the PC. However, the meaning of emotion / emoção is stable in that it refers largely to generic, undefined, or undefinable emotional circumstances, or to mixed emotions. Although, given a large enough context, one may not need to be a psychologist to deduce the type of emotion felt, emotion is often used because the SENSER, or the Speaker, is uncertain how to interpret the general emotional situation.

The most popular lexeme in the EC is feel, 51,5%, and, in the PC, it is sentir-se, with 55,7%. Their more central meanings are fairly similar, but they differ as to their more peripheral meanings, and, consequently, as to the type of complementation used. Feel has a high proportion of more consciously recognized examples, as can be seen by the large number that are complemented by either (THAT) or comparative clauses, 37,5%, whereas similar structures with sentir-se are much less frequent, 10,9%. The Portuguese translation of many of the English (THAT) clauses would more easily accept achar (similar to think/find) than sentir as an introductory verb.

In the physical sense, feel is usually restricted to perception through touch, and visceral, or internal, reactions. However, sentir-se also covers areas which would normally be translated into English by hear, smell and, sometimes, see. Although Portuguese has perfectly valid specific lexemes for these situations - ouvir, cheirar and ver - , it uses sentir to express an almost pre-conscious perception of certain events, as can be seen in the following example:

(13.3) O volume daquele som, que se comunicava com o de milhões de pequenos crótalos de prata que ao mesmo tempo se chocavam entre si pelo efeito da própria vibração, enchia todo o vale, ocupava todo o espaço, como um sólido. Sentia-se, sem se ouvir. Si 200 (My italics)

A similar meaning can be expressed by the English sense, but the examples are relatively rare.

If one is translating the more psychological uses of the noun feeling one will probably use sentimento. It is the second most popular item in the PC group, accounting for 11%, and it only specifies the emotion in the syntax of 10% of the examples, a percentage just slightly lower than that for specified examples of feeling, 14,6%, and even at the pragmatic level, only about a third of the examples for both lexemes specify the PHENOMENON in the context. The specific examples include emotional situations like um sentimento fundo, de irreprimível repulsa, but there are more complex examples such as sentimento de justiça. One usage, as a 'social formula', is the phrase os meus sentimentos for funerals, similar in meaning to the English condolences. The English cognate, sentiment, is infrequently used in modern English, and tends to be rather formal, as in:

(13.4) He preserved from collectivisation some discernible part of his personality ; he could not be drawn at moments of sentiment to talk of his girl, his family or his children. SPY

The adjective sentimental, in both languages, is more restricted in meaning than the noun and is usually used to describe situations of affection. However, in English, its use often implies a negative assessment of the SENSER's behaviour as showing weakness of character or even silliness.

The next most popular noun in this group in English, sense, which often refers to the functions of perception, as qualified by sight, smell or feeling. It can be used generically, particularly in the plural, but all but 13,4%(E) of the 112 examples were complemented by specific nouns. However, although some of these nouns were Emotion words like joy and gratitude, others were more intellectually orientated, as with beauty, decency and innocence. Sense may be used as a SFoc verb, with a meaning comparable to a more psychological or intuitive interpretation of feel, as in:

(13.5) At that moment Karl seemed to hear some sound, sense danger; he glanced over his shoulder, began to pedal furiously, bending low over the handlebars. SPY

In Portuguese, the noun sentido, usually in the plural, is used to refer to the functions of perception which, in English are called the senses. However, that is as near as it gets to the semantic field of Emotion, being more frequently translatable by sense, when this word implies 'meaning' or 'direction'. As a past participle of sentir-se, it can be found on rare occasions, as in;

Sensação largely appears in examples where sensation, or the more physical interpretation of feeling, would be an appropriate translation. However, it can also be used with more psychological examples, such as sensação de perigo/tristeza, which in English would be rendered with sense or feeling.

Passion and paixão can both be used generically, but they more often appear in contexts where their meaning is specific, and care should be taken when translating them. The 28 examples of passion in the EC divided themselves fairly equally between the generic sense, and the meanings of the Love or Desire and Anger groups, but out of the 141 examples found for paixão, 83% had to be classified as strong lexemes in the Liking group, and the remainder were generic. When used generically, both words can appear either in juxtaposition with emotion/emoçaõ, or in expressions like paixões humanas or the passions of this earth which are arguably antiquated versions of the more modern uses of emotion.

The less central items in this group have been included to show how emotion is related to mood and temperament. Also, when translating, these peripheral concepts can be confused because of their superficial resemblance as cognates, so it is as well to distinguish between them. For example, disposition and disposição appear to be cognates but, disposição or disposto are far more frequently used than disposition and disposed. This is probably partly because the English lexemes are considered a little old-fashioned, and partly because temper and (good / bad) tempered would be more suitable modern translations of disposição and disposto in many cases. The few examples of disposed could be translated by disposto, but disposition would be better rendered by temperamento. Disposição, on the other hand, is best translated by mood.

Mood is used to refer to a person's current state of mind, as it is affected by the emotions, but only a third of the examples were used in a truly generic sense. A few examples, when the context is analysed, refer to positive situations of a less emotional kind, such as this mood of old friendship and autumnal mood. However, more examples refer to negative situations, as in her moods of irritable depression and a mood of vehement self-reproach, and the adjective moody and its corresponding adverb, moodily are used only in negative situations

Humour, in its generic sense, is positively orientated, often qualified by good, and the adjectival form is good-humoured.. The Portuguese humor, however, is wider in meaning, covering generic, negative and positive situations, and being equally easily qualified by bom or mau. For this reason the negative situations are better interpreted by mood or temper in English.

Both humour and humor can refer to situations describable as amusement, and the link between the two words, particularly in the case of humour, is interesting. One could argue that being amused, or finding something funny, involves a spontaneous, and sometimes uncontrollable, reaction which is classifiable in the Joy group of Emotion. Perhaps the reason why amusement is not normally considered an Emotion lexeme is because the situation it describes would seem to arise from cognitive factors traceable to the situational, social and cultural background of the person involved. However, if, as some believe, emotion involves more cognitive appraisal than was previously supposed, then one could argue a case for reactions caused by this meaning of humour/ humor to be classified as quasi-emotional, or at least as a causative factor of Joy.

As one moves away from the more central lexemes one comes to those like sensibility, sensitivity and nerves in English, and their cognates in Portuguese. Some of these cognates, however, are false. Sensibility is restricted to the more psychological capacity to feel Emotion, but the adjective sensible, in modern English, is related not to sensibility but to sense as in good sense or reason. Sensitivity and sensitive, on the other hand, cover everything from a capacity for understanding or emotional feeling, through physical feeling, to the behaviour of exceptionally well-tuned machinery.

The modern English sensible is usually translatable by sensato in Portuguese, but sensibilidade and sensível have the wider applications of sensitivity and sensitive. Sensitividade and sensitivo would seem to be restricted to the more physical senses and rarely used directly of a person. Nerves and nervos both refer to a general pre-disposition to irritation, anger or fear, but the singular form in English, nerve, seems to be related to courage. The adjectives nervous and nervoso were considered in the context of the Fear group.

Although rather peripheral to Emotion, these lexemes are interesting insofar as they focus both the physical and psychological tendencies of the SENSER to be affected by events which cause emotion. These lexical items show how, for several centuries, folk wisdom has reflected in language the belief that the senses / sentidos and the nerves / nervos, whatever the current interpretation of these terms, form a sort of interface between the physical and psychological areas of our perception.

Difficulties in classifying the Generic examples according to PHENOMENON type

Attempts were made to analyse this group using the PHENOMENON types proposed, but they proved rather fruitless, for various reasons. Some of the lexemes, being generic in nature, had been used in the first place to avoid specification of the emotional situation, as in:

(13.7) "You shut up!" she shouted, in a blast upon the mottled majesty of the old lady. The old woman's breast heaved with heaven knows what emotions. VG

or because the Speaker was unsure how to interprete what was happening to the SENSER. At other times generic words combine with an Emotion word, as if to emphasise and specify it, as in:

(13.8) He told me he was experiencing a feeling of unutterable relief, of vengeful elation. LJ

One could find a PHENOMENON for both of these examples in the context, but examples such as:

in which the emotion and the PHENOMENON are both clearly stated are not as frequent as one might suppose. Apart from the fact that notions other than emotion were often involved, the reason why it was difficult to assess the underlying PHENOMENON of the more emotional ones was because of the complexity of the physical, emotional and cognitive factors at work, and two or more processes could often be distinguished as PHENOMENA. For instance if one says I feel miserable, this may be induced by several factors, i.e, I am cold and wet (SENSER's physical state > PH. type 3), My boyfriend has left me (affective problem caused by other's behaviour > PH. type 9) and I realize that I have three miles to walk in the rain (intellectually assessed factor in my world > PH. type 11). And this is a simplistic diagnosis, as most emotions can be seen as due to our psychologists' constellations of stimuli, and this, in effect, tended to lead to a lot of examples being classifed as PH. type 1.

The principal syntactic feature to be noticed in this group is the special behaviour of the verbs feel and sentir-se. Besides being the most popular lexemes, they are significant because they do not behave like normal SFoc verbs. They can be used as copula type verbs in SPC structures, and as transitive verbs in SPO structures, and sentir-se behaves in other interesting ways.

Both feel and sentir-se act as copula-type verbs which can be complemented by nouns, adjectives, past participles and adverbs, as in:

In this sense they obviously focus the Subject because both the syntactic subject and the syntactic complement refer to the Subject. Nearly 40% of the examples for feel were of this type, but the same was true of only about 28% for sentir-se. Most of these examples took adjective / participle complementation, the EC producing practically no examples of noun complements, and the PC only 14 examples, some of which are arguably adjectival because the noun has no determiner, as in Ele se sentia homem. 26% of the examples with feel took an adjectival complement, but the past participle complement, with 12,5%, was the most popular with sentir-se.

When feel and sentir-se behave like copulas, it is arguable that, since it is the complement which defines the more generic meaning of the copula, the copula itself is little more than a space-filler. When we say He is happy, for example, the verb be merely marks a general idea of state, and with He feels happy we have a rather more specific notion of Emotion which is also considered a state by such as Quirk at al (1985). There seems to be little to dispute here as far as the majority of examples with feel is concerned. It would seem sensible, also, to accept the analysis of sentir-se as a copula which, like other verbs described by Vilela (1992 : 77-9), has simply incorporated the reflexive -SE as a lexical, rather than syntactic, part of the verb and is an inherent reflexive. However, I should like to argue that, for sentir-se, one can find a semantic gradient between a structure which expresses a simple psychological state, for which the above syntactic explanations are perfectly adequate, and a more process-like, reflexive structure which might account for certain uses of sentir-se and several examples of feel + -SELF.

If we wish to paraphrase John is happy, it is difficult to see how we can do so unless we use Wierzbicka's system[1]. However, if we want to paraphrase John feels happy, we can easily say John feels he [John] is happy with only a minimal shift in meaning to the idea that John can contemplate his Self, or whichever term one prefers, as being happy. This shift can also be expressed in Portuguese, and it is significant that one has to drop the -SE of Ele sente-se feliz to produce Ele sente que [ele] está feliz. In this case, the extended sentence is an SPO sentence, the (THAT) or QUE clause acting as Od. Although the EC produced only one example of feel + -SELF with an adjective:

(13.11) His year of anarchy had filled a deep, interior need of his, the escape from reality, and as he found himself increasingly hemmed in, where he once felt himself free, he became at times listless and morose, even with me. BH

(13.12) Herbert had felt himself obliged to confide the state of the case to me. GE

(13.14) 'No one,' I affirmed, feeling myself swayed by some strange excitement. LJ

(13.28) He took off from the last dry spot, felt himself flying through the air, felt himself without any shock, planted upright in an extremely soft and muddy mudbank. LJ

In the examples with adjectives, participles and adverbials, the -SELF pronoun can be omitted without changing the syntax of the sentence, but it is essential with the -ING clause structures.

The Portuguese corpus also produced a few examples in which sentir-se was complemented by an infinitive clause, as in:

The English -ING participle clause, described in detail in Quirk et al. (1985:16.42/53/83), does not have an equivalent -NDO construction in Portuguese, but is usually translatable by an infinitive clause. One can therefore suggest, for the present situation, that the English -ING clause after feel + -SELF, and the Portuguese infinitive clause after sentir-se, are parallel structures. Therefore, one can further suggest that the -SELF / -SE particles are performing some kind of syntactic function as S in relation to the subordinate clause[2]. It can, therefore, be interpreted as semantically reflexive because it refers to the S of the sentence which also happens to be the S of the main clause.

One could simply dismiss the examples with the optional -SELF as examples of emphatic usage, or explain them away as vestigial from some older structure, or even argue that they are signs of an emerging Self consciousness. If there were no structure in Portuguese, and other languages, with a -SE pronoun to point to, one could, no doubt, enjoy speculating. However, the fact that the -SE pronoun exists does allow one to postulate a psychological and semantic degree of reflexiveness for the feel + -SELF and sentir-se examples, even if they cannot be considered true syntactic reflexives. In other words, these examples with sentir can be seen as somewhere on the gradient described for irritar and zangar-se, as discussed in Chapter 7. This interpretation could still use the term of 'reflexivos inerentes', or inherent reflexives, to distinguish these situations from the wash oneself or lavar-se, or true reflexives. However, instead of using it as a syntactic term to explain the lexical nature of the -SE, one could use it as a semantic term to explain the psychological process inherent in the use of -SE. If, like Halliday (1985 : 111) and the Cognitive psychologists, one would prefer to consider Emotion a cognitive process rather than a passive state, this interpretation is semantically sound.

The examples in English lend themselves to interpretation as SPOC structures, similar to the non-reflexive one below:

(13.21) And here she was, she reflected, feeling life rather sinister again, making Minta marry Paul Rayley; VW

Although not very frequent in either corpus, these examples, and parallel ones in the PC, like:

[13.22) Cada vez amava mais o que é pobre e o que é fraco. Em Santa Olávia, as crianças corriam para eIe, dos portais, sentindo-o acariciador e paciente. M 1

As we saw in Chapter 6, the SPOC and SPOA structures are related to the simpler SPO structure on a gradient, and the question, in the PC, is whether to grade SENTIR-SE as a lexically extended copula and call all these sentences either SPC or SPA sentences, or whether they qualify as SPOA and SPOC sentences, structurally related to the SPO of the extended sentence type John feels he is happy. Although admitting that there is a gradient of acceptability here, I have opted for the SPOC / SPOA analysis.

One can use feel and sentir to indicate the simple idea of physically feeling, or sensing, without any PHENOMENON being implied. Examples of this are unusual and the EC produced none for feel. The PC produced a few infinitives of this kind and one or two examples like:

The Portuguese Passive -SE construction, translatable in this case by either One could feel X, or X + was felt, accounts for 4,5% of the examples with sentir, most of them referring to physical PHENOMENA. It is interesting that the few examples like this in English either appeared as passives with the PHENOMENON acting as S, as in :

(13.23) I was about to ask him what he meant when a sort of preparatory tremor passed over his whole person, as a faint ripple may be seen upon stagnant water even before the wind is felt. LJ

(13.24) The air felt cold upon the river, but it was a bright day, and the sunshine was very cheering. GE

There are very few examples on which to base any definitive ideas on this subject in English, and I see no reason to contest the Passive interpretation of the Portuguese -SE structure in these cases.«

The most popular structure in the PC was sentir, without -SE, followed by a simple noun as direct object. The more physically orientated examples accounted for 13% of the examples and the more psychological ones about for 44%, making a total of 57%. However, feel produced a total of only 22%, 9,5% physical and 12,5% psychological. In both languages these objects were often Emotion nouns, and the more physical examples frequently referred to the body or its parts.

The nouns used are often of a directly emotional nature, such as distress, jealousy and pity, or alegria, desgosto and gratidão. However, there were others of a more intellectual variety, particularly with feel where examples such as injustice and a sense of responsibility were fairly easy to find. Sentir provided few such examples, and these, too, had an emotional interpretation, such as a sua importância no mundo and uma fraterna tolerância. Sentir, on the other hand, seemed to favour nouns which referred to the causative aspect of the PHENOMENON, such as a ameaça duma ruptura, and the notion of how the SENSER contributes to his/her own emotional situation, as with a sua incompetência, an example paralled by my deficiencies with feel, although this is a rare use for feel.

One of the most noticeable features of the nouns in this group is the fact that they are often overtly countable. This is more noticeable in the EC, which has the lower percentage of countable nouns with the specific lexemes, 13,6%(E) as against 20,8%(P), because the relative situation is reversed with 54,1%(E) and 41%(P). The non-plural/singular group accounts for similar percentages, 38,8%(E) and 37,4%(P), very similar to the overall average found for both corpora with the specific lexemes. This leaves the non-countable group of generic lexemes with 8,5%(E) and 20,2%(P). Thus, although both corpora show a tendency to count the generic lexemes more than the specific ones, the difference between the EC and the PC is marked by the reversal of positions over countability with generic in relation to specific lexemes. Moreover, those examples with definite articles and possessive pronouns favour analysis as instances or relations, rather than a mass interpretation, when examined in context.

The singular usage obviously occurs with expression like a feeling / sense of X, or um sentimento / uma sensação de X, so expressions like this, using a definite article instead of an indefinite one, can be fairly safely interpreted as singular too. The plural usage is particularly marked by its Nominal type reference, referring to either the concept in general, or to someone's range of emotional capacities. This is particularly true of the use of feelings, of which only 37,6% were specified in the co-text. Some of these were specified by Emotion words like pity or relief, others by more physically connotated adjectives like warm and sexual. Both emotions and emoções, behaved in a similarly general way, and so did senses and sentidos, and sentiments and sentimentos. The non-countable examples tended to occur in BEHAVIOUR situations, in expressions like his eyes, glazed with emotion or aquele ar de sentimento e de poesia.

The first thing to notice about the adjectives, participles and adverbs is that, even added all together, they account for few of the examples in the generic groups, 11,4%(E) and 10,4%(P), and most of them are associated with the less central lexemes. The adjectives and adverbs in the EC largely refer to behaviour with passionate/ly, and a large proportion of those in the PC is taken up with disposto. Otherwise the examples with moved / touched and comovido match each other, and there are others with sensitive and sensível. In both corpora, sentimental has a largely PFoc meaning.

Perhaps the most significant point to be made about these lexemes is that, although they primarily fit into the category of Emotion, they also cover all the mental processes of perception, cognition and affection described by Halliday (1985 : 111), and the states described by Quirk et al. (1985 : 4.29) as intellectual, and of emotion/attitude, perception and bodily sensation. This is particularly true of feel and sentir, as we have seen. Both appear to be complemented by lexemes of physical, emotional and intellectual types, although to varying degrees, and feel was often complemented by (THAT) clauses, in a way which could be substituted by think, except that feel would suggest intuitiveness, and think reason.

The same problem, of course, also applied to the specific Emotion lexemes, but to a lesser degree. This was partly because the generic lexemes usually require further definition by the specific ones, and thus function at one remove from the actual emotion. Linguistically the specific lexemes allow one to distinguish the immediate PHENOMENA more easily and definitely.

There is another interesting point, which may represent a compensating mechanism between the languages, but also suggests that the EC shows a tendency to favour the distancing, or indefinite, factor of the generic lexeme more than the PC. The number of specific examples extracted from the corpora shows the PC with a higher incidence of specific emotion words, 1,47%(P) of all words in the texts, as against 1,24%(E), a variation of 0,23% (or an 8,5% difference) in favour of the PC. Generic examples, however, account for 0,126%(E) and 0,132%(P), a variation of only 0,006% (or a 2,3% difference). One may deduce from this that, although the EC still shows less overall interest in Emotion than other lexical fields, it seems to employ generic lexemes relatively more frequently than the PC.

The most important lexemes in the corpora are the cognates surprise and surpreender, with 48,5% and 35,4% respectively. Lexically they are very similar, although their syntax varies more. After that, the EC is largely composed of examples of amaze (19,3%), startle (12,4%) and astonish (12,1%), the remaining 6 lexemes accounting for only 7,7% between them.

As usual the lexemes are more evenly distributed in the PC, with espantar (23,9%), pasmar (13,4%), assombrar (11,3%), admirar (9,5%), with 5 lexemes accounting for the remaining 6,5%. All the less popular lexemes in both corpora are used as more extreme expressions of Surprise. However, it is difficult to draw easy correlations between the meanings of the other lexemes, except to say that one roughly grade startle > astonish > amaze and admirar > espantar > assombrar in ascending degrees of strength. However, startle and assombrar contain traces of Fear not present in the others. Pasmar is unusual in that it is more related to behaviour and the type of facial expression best translated by phrases like he gaped in amazement or his jaw dropped in surprise.

As can be seen in Table 13.1, this is an area in which an exterior PHENOMENON is identified in 82,9%(E) and 86,6%(P) of the cases. PH. type 1 examples are very few, although the PC registers some which result from multiple phenomena. The EC has 16,3% in the 3-5 band, and the PC 7,8%, but type 4 is predominant. In the 6-9 band, type 9 is the most numerous. None of the lexemes stands out as being different from the general tendencies shown in the overall picture.

The SFoc/PFoc ratios of the two corpora are fairly similar, and the same is true of the number tagged for BEHAVIOUR. However, the way in which different parts of speech assume the PFoc or SFoc roles varies considerably.The SFoc items show a similarity, for once, in the percentage of noun forms. Countability is low in this group, being almost non-existent in the EC, and both corpora showing a particularly high level of explicitly uncountable nouns. Complementation is virtually non-existent too, but there are an exceptionally high number of adverbial phrases.

There is also a high number of SFoc adjectives and past participles, the latter alone accounting for 41%(E) and 26,5%(P) of all examples. The EC adjectives are few, and consist of unusual words like agog and dumbfounded, and most of the PC ones are of atónito. The participles are largely represented by amazed, astonished, astounded, startled, stunned and, above all, surprised, in the EC and by admirado, assombrado, espantado, pasmado and surpreendido, in the PC. The participles are complemented in 53,3%(E) and 29% (P) of the cases. The number of zero copulas in the PC is lower than usual and there is the high percentage of 25,2% with ficar, a small number of examples with both ser and estar, and a certain variety with other quasi-copulas like parecer. The copula situation in the EC is not unusual.

The big difference between the corpora is the fact that whereas the EC has no SFoc verbs the PC boasts 9,1% of them. A few of these are of admirar + clausal complementation, but most of them are with pasmar. The emotion features of this verb are undoubtedly there, but there is perhaps a stronger element of the behaviour associated with it.

The PFoc adjectives + adverbs are the dominant pattern in this area of the EC, accounting for about two-thirds of the PFoc lexemes, most of them occurring with amazing/ly, astonishing/ly, startlingly and surprisingly . However, the adjective and adverb forms are rare in the PC, being restricted to assombrosa/mente, espantosa/mente and surpreendente /mente.

Although there is no ambivalence in the adjectival forms, the noun forms which appear as PFoc in both corpora are all ambivalent. This is particularly noticeable with the larger amount of examples in the PC where assombro, espanto, pasmo and surpresa are all ambivalent on an overall ratio of 69/64. Only surprise is involved in the EC with a 16/50 ratio. The EC seems to show a pragmatic preference for phrases like It is surprising over It was a surprise, and the PC prefers the opposite structure, Foi uma surpresa over Foi surpreendente.

The percentage of PFoc verbs is higher in the PC (19,1%) than the EC (8,3%) and the ambivalence of the nouns is reflected in some unusual uses of these verbs. Although in both corpora the main structure is a simple transitive verb with an 'a' type object, 60%(E) and 60,7%(P), the EC has 9 examples in which the existential IT is used, as in It did not surprise me to find Mr Samgrass, and the PC produced 5 examples of apparently intransitive examples like Um pouco de sorte e a Virgem do Pilar por seu lado, e o milagre não era coisa de espantar. (AQ) There were also 29 examples with -SE of a semi-reflexive type. Many of them were with admirar and the ambivalence of this verb is enhanced by the fact that it can function both as an PFoc and as a SFoc one with the meaning of Surprise. The classification of the -SE examples as either PFoc or SFoc is therefore rather arbitrary, although the decision was taken because of the large number of intransitive type ones which suggested that the interpretation of the Self as PHENOMENON was possibly quite valid.

This group has several syntactic features in common with the other Emotion lexemes. PFoc participles and adjectives describe the qualities of the PHENOMENON which cause the emotion in the SENSER, and there is the familiar PFoc verb + SFoc past participle combination. There are several cases of ambivalence among the nouns and verbs which indicate the symbiotic relationship between SENSER and PHENOMENON, as well as of -SE verbs which seem to fit with difficulty into either the PFoc or SFoc category. So one can say that, generally speaking, the syntactic behaviour of this area could justify its inclusion among the more central type emotions.

However, the high specificity of the PHENOMENON types belies this apparent centrality. The more central emotions with a similar syntactic profile, like Joy and Distress, have high levels of PH. types 1-5, but Surprise clearly specifies the PHENOMENON, favouring types 9-11, and even the few in band 3-5 tend to be well-specified in context. On the other hand, the fact that the Surprise emotion is easily identifiable from facial expression is confirmed by the high numbers of examples tagged for BEHAVIOUR.

So in a way everyone is right. The emotion theorists who insist on physical symptons as a proof of emotion are right in accepting Surprise as an emotion. The linguists can point to the syntactic behaviour to justify the description of these lexemes as emotion because of their syntactic patterns. Those who prefer to put emotion down to complex cognitive processes can point to the specificity of the PHENOMENA, and say that it is too cognitive to be emotional. And this brings us back to the problem we have been trying to solve since the beginning - to what extent can the emotions be described as cognitive processes.

Desire is not considered to be an emotion by Ortony et al., and they see the desirability or otherwise of something as a variable affecting other emotions. Those who need a universally recognized facial expression of an emotion to justify its inclusion in the list also reject it. Izard (1991) describes facial expressions, which he relates to Interest, which might be understood as expressing the early stages of what is eventually realised as Desire, but he does not consider Desire itself as an emotion. The only list of basic emotions quoted by Ortony et al. (ibid: 27) which includes Desire is that of Arnold (1960). Fridja (1986) also considers that it should not be ignored.

There is a certain element of emotion in Desire in that it expresses a gut reaction rather than a reasoned approach to something. Whether it is sexual desire, a craving for drugs, or the longing of a child for a particular new toy, it is difficult to deny the physical and psychological tension that accompanies it, or the fact that it can lead people to such anti-social behaviour as rape, theft, and other crimes, in order to achieve satisfaction. Although those who choose emotions on the basis of universally observable facial expression do not recognize Desire specifically, there is no doubt that we recognize certain physical indications of this emotion, although they may be part of the more generic symptoms of excitement, expectancy or interest, or be described semi-metaphorically in terms related to the physical appetites of hunger, thirst or sex.

If one were to attempt a definition of Desire, one could describe it as prospect based because the time dimension is definitely focused on the future. In Ortony et al.'s type of classification it would involve EVENTS, AGENTS and OBJECTS, because it can be used with things that happen, the people who cause things to happen, including the SENSER, and a wide variety of objects.

Negation of the lexemes under consideration was frequent, particularly with want, but these examples were not analysed because they normally only focused the simple negation of desire or volition to do something[3]. These examples would have to be examined if the implications of not wanting were to be considered, but I feel it is unnecessary here to examine the non-existence of something which is not even considered a full emotion by most authorities anyhow.

Although there are several lexemes which can indicate various degrees of Desire, this group is dominated by want (65,5%) and wish (19,2%) in the EC, and querer (77,9%) in the PC. Desire and desejar only account for 4,9% and 12,8% of the group in their respective corpora. The only other lexemes of any numerical interest are eager (3,6%), inclination (2,1%), and long for (1,9%), in the EC, and apetecer (3,3%), ávido (1,1%), cobiça (1,6%) and sôfrego (1,9%) in the PC.

Want and querer are near-equivalents, particularly in the most usual sense when they imply volition as much as desire. They differ in their more marginal senses. Want, with a person as syntactic Object, is usually a blatant expression of sexual desire, but although querer, used in the same way, can mean much the same, it tends to be softer and more related to love.

Wish is unusual in that it implies Desire about a hypothetical future and, and takes either the conditional, which tends to express a possibility that the desire may be satisfied, or the subjunctive, which normally implies a near impossible desire. Thus we get, for example:

These sentences are difficult to translate as there is no verb in Portuguese which behaves quite like wish . Only the context can really tell us whether one needs to translate the first example by:

which express stronger or weaker approximations to the original, but not direct translations. The second example is even more difficult to render, and one can only do interpretative translations like:

Desire and desejo are fairly similar when used in a strong sense. The EC examples use desire to express a strong feeling or wish, but it is no longer used in the weaker, more polite, type of phrase, like What do you desire? which is decidedly old-fashioned. However, the Portuguese question O que é que (a senhora) deseja? is a perfectly normal, polite, modern way of saying What do you want?, and is, for example, the cultural equivalent of the What can I do to help you, madam? that a superior salesperson might use to a customer in a shop.

The other lexemes describe different ways of Desiring. Apart from those already mentioned, the more central ones, are hanker and long for, and anelar and ansiar, which all express a rather more long-standing desire and imply little hope that the desire will be satisfied. The weaker type of Desire lexeme, like care, as in:

(13.28) 'Why, don't think about it, Mr Ryder. It was a pleasure,' he said, 'but anything you care to give is useful in a parish like mine.' BH

(13.29) Without evincing any inclination to come in again, he there delivered his valedictory remarks. GE

(13.30) 'I prefer not to anticipate my communication here; you will impart as much or as little of it as you please to your friends afterwards; I have nothing to do with that.' GE

and apetecer, are useful in situations where it is of social interest to play down too much interest. There are then two groups which are used in quite strong contexts, to refer either to desires which are seen as positive in nature, as with eager and zeal; or the more negative types of Desire, as with avid, covet, crave, greed and lust, and ávido, cobiçar, gana(s) and sôfrego.

The influence of want, wish and querer control the results here. The nature of Desire, and of these verbs in particular, would seem to require that the Object, or PHENOMENON, should always be explicit. Yet there are a small number of PH. type 1 examples, as can be seen in Table 13.2, in both corpora, largely of the What do you want?, or ..será o que Deus quiser type, in which the nature of the PHENOMENON is, by definition, unknown. Others are found with adjectives like eager, ávido and sôfrego in which the actual PHENOMENON is ill-defined. With desejo, the examples are often found in a generalized plural, as in:

(13.31) - Quando me ponho a pensar no que era a minha vida, os meus desejos, as minhas ambições, fico um bocado desconsolada. SU

Otherwise there is a strong tendency towards types 3-5 in both corpora, 48,9%(E) and 57,9%(P), of which most are type 5, to be found with care to, desire, inclination to, long to, want and wish in the EC, and with ansiar, apetecer, desejar and querer in the PC. The PHENOMENON types here match the use of 'b' clauses in the syntax of these examples.

Reference to the Other as the PH. type 6 is found in 4,3% of the examples in both corpora, although these examples are rarely of the kind that imply sexual desire. The difference between the two corpora is to be found in the fact that the EC gives us 19,4% in types 7-9 whereas the PC gives us only 6,1%. However, to balance this the type 11 examples in the EC are far fewer than in the PC. This reflects the fact that the syntax of want, and some examples of wish, requires a 'c' clause and the reference to the Other's actions is perfectly clear, as in:

The infinitive clause may not be as positive as the normal active voice, but it is not as hypothetical as a conditional or a subjunctive.

With querer, however, one has to to use 'g' clauses, usually with the Conjuntivo form of the verb in the subordinate clause, and this tends to make it more difficult to distinguish between types 9 and 11. Although type 9 can be attributed quite easily to:

because it refers to a definite person coming to a definite place, despite the hypothetical element introduced by the Conjuntivo, and type 11 to:

because the reference is to the hypothetical fact of whether or not it will rain on Sunday, examples such as:

is more difficult to assess as type 9, as no specific person is referred to, merely the act that someone unspecified might perform, making this example more appropriately described as type 11. Similarly, although it might be possible to classify:

as type 7, in which the Other's state is focused, type 11 is perhaps a better interpretation because the overall state of affairs is not simply the Other's state, but the hoped-for fact of the Other belonging to the SENSER. There are several rather fuzzy examples here, and the difference is caused by the syntax which makes it necessary to be more fact-like with querer, and more Other-focused with want. I do not pretend to have solved them all satisfactorily.

The PH. type 10 examples, in which a non-human object is desired appear in a similar ratio in the corpora. They are not difficult to identify, and it is interesting to see the similarity of the data in the two corpora for these and for those of type 6.

One of the most noticeable things about this group is that it is nearly entirely SFoc, and there are few examples tagged for behaviour. The group was not considered in the overall corpora averages because of the marginal nature of Desire in relation to Emotion but, if it had been included, the data would have been distorted because it would have accounted for 13,8%(E) and 14,8%(P) of all the examples in the corpora

The SFoc verbs are the most important feature of this group, and if these verbs had been included in the overall data for the corpora, the Desire verbs would have made up 42%(E) and 50%(P) of all the SFoc verbs Although there are a few examples in the EC with care to, covet, crave, desire, hanker for, incline to, long for and please, the vast majority of the examples are of want and wish, with 966 and 264 examples respectively. Only 32,3% of the examples with want take simple direct objects, with a further 1,9% of the SPOC type. The rest take clausal complementation, 53,3% with 'b' clauses, and 12,5% with 'c' clauses, and nearly all the examples of wish take clausal complementation, 30% taking 'b' clauses, 8% taking 'c' clauses, 18,6% taking 'f' clauses, and 35,6% taking 'g' clauses,. There were also a few SPOA and SPOO examples.

In the PC, anelar, ansiar and cobiçar produced a few examples but most of the SFoc verbs were accounted for by desejar with 123 examples and querer with 1592. With desejar 38,2% of the examples took simple noun phrase objects, 43% 'b' clauses, 10% 'g' clauses with the remaining 8,9% being accounted for by a few intransitive examples, -SE pronouns, and prepositional objects. Querer took 19,5% type 'a' objects, 62,8% 'b' clauses, 16,6% 'h' clauses and 1,1% intransitive and -SE examples.

The SFoc nouns are most numerous are with desire, wish and want, with a few examples of eagerness, hankering, inclination, lust, yearning and zeal. In the PC, desejo accounts for 67,8% of the nouns with few examples of nearly all the other lexemes. In both corpora the tendency is towards countability, a high level of complementation, but few adverbial noun phrases.

The few SFoc adjectives are found with desirous, eager, greedy, wishful and zealous in the EC, and ávido, cobiçoso, desejoso and sôfrego in the PC Most of the examples in the EC take complementation and over half of those in the PC do too. For such a small number of examples the EC shows a fairly wide selection of copulas but most of the PC ones take the usual zero copula.

The few PFoc lexemes in the EC consist of examples of the adjective desirable and some marginal past participle, or passive examples of want. The larger number of PFoc examples in the PC is largely due to various forms of apetecer, but this is largely because of the unusual behaviour of the verb. Although the PHENOMENON is the Subject of the verb, it rarely appears before the verb. Instead, it appears after the SENSER which, in its turn, nearly always appears in a dative form. There are also a few past participle examples with cobiçado and desejado, and a few ambivalent adjectives like ávido and sôfrego.

If one turns to language usage for clues that might help to solve the problem of whether Desire can be considered an emotion, the most significant aspect to be analysed is the high proportion of SFoc verbs. As we have seen with other groups, like Liking, the use of Active forms of SFoc verbs is a sign of a high degree of conscious evaluation of the emotional processes, and demonstrates positive affirmation of this evaluation. In this sense, the use of a SFoc verb can be seen as usable only when the emotional process has completed the 'appraisal' phase and is at that of 'action readiness'.

The analysis of the PH. types shows up a variety of objects of Desire, from simple noun phrases to complex propositions about the world. This does not necessarily mean that the simpler PHENOMENA are more directly the result of emotional processses than the more complex ones. The emotional strength of I want a drink, I want to pass my exams, or I want my son to pass his exams can only be measured objectively in context. The simplicity of the PHENOMENON merely indicates a simpler cognitive process behind it.

One might well argue that Desire develops from the Appreciation and Liking emotions, because it can often be seen as a consequence of them. Similarly, if one were considering the negative uses of Desire lexemes, and the use of lexemes such as reluctance, one could argue that they are the consequences, in part, of the Dislike, or even Distress or Fear emotions. Yet Ortony et Al. see desirability as a variable which influences the Reaction to Events emotions, and this would suggest that the evaluation of desirability, or even of Desire, about the event, precedes rather than succeeds, these emotions. However, the need to argue for such a linear succession of events is not so pressing if one remembers the points that were made in Chapter 1 about the complex processes in the brain, which allow for raw perceptual input to both influence and be influenced by feedback from all the previously stored information in the brain.

Naturally, a girl will be happy if she has a telephone call from a boy only if she wanted a call from that boy. Furthermore, the desire to receive the call may depend on her liking the boy, but it may arise from some quite unrelated fact like some happy-making information she needs which only he can give her, in which case the liking factor is directed at the information and not at the boy. In a situation like this one has an apparently linear chain of events which proceeds from liking to desiring, to receiving and to being happy. One could even continue the chain of events, and say that as the girl was happy, she wanted to celebrate, or she liked the suggestion that they should go to the cinema tomorrow, or various other possibilities which a happy frame of mind can provide.

The net result of this example is to draw attention to the way our emotions are interlinked and to suggest that rather than set linear sequences of emotions, it is more appropriate to talk of the sort of interrelating networks which experts now feel are more suitable for human thought processes. In any case, as far as Desire is concerned, whether one considers it simply as a variable or as an emotion proper, it would seem to be caught up somewhere in the overall emotional process. Psychoanalysis has shown that desires are often unconscious or subconscious, and under these conditions their contribution to emotion is complex and subtle. However, by the time desires are actually expressed in language, their nature is clearly recognized.

[1] Wierzbicka (1992) in Cognition and Emotion describes this type of situation thus:

X is happy

X feels something

sometimes people think like this:

something good happened to me

I wanted this

I don't want other things

because of this, they feel something good

X feels like this

[2] The Portuguese examples could, in fact, be modified to allow a different subject:

........sentindo o João triunfar.....

Elias sentiu a Maria a gelar perante tanto fatalismo.

This, however, is possibly due rather to the way sentir, in these examples, is more like the intuitive sense than feel. The English examples are distinctly related to physical feeling, and the -SELF particle can only really be substituted by his body in these examples. One could, however, construct more physical examples with sentir, and more psychological examples with feel.

[3] The lexemes reluctance and relutância were collected but, although they express a definite sense of 'not wanting' and can be found with the copula feel, they do not really fit into this group, and to include them would have meant considering other lexemes such as hesitant.