A Period of Retrogression

 

• Hess-Lüttich & Niederhauser 1995. Textlinguistics: Coherence, reference, and cognition
• Power & Dal Martello 1985. Methods of investigating conversation
• Makkai 1987. Idiomaticity and phraseology in post-Chomskian linguistics

---

Hess-Lüttich, Ernest W. B.; Niederhauser, Jürg 1995. Textlinguistics: Coherence, reference, and cognition. Semiotica 107(3/4): 349-354.

'Text linguistics' is a relatively new field in the long history of linguistics. It has, however, developed rapidly over the last 25 years or so. Today, text linguistics is a well-established sector of general linguistics. In Germany, for instance, two of the leading journals have just devoted special issues to the topic (cf. Brinker 1991; Klein 1992). These volumes give a reasonably good overview of the various developments in the field and represent a wide range of contemporary trends. More specifically, they aim at bringing together various approaches which up to now have been followed quite independently from each other:

• the structural approach focusing on 'text' as a 'product', as a series of units of lower orders, of phrases, words, morphemes, letters and the rules for connecting them (cf. Harweg 1968)
• the pragmatic approach considering texts as comunicative units functioning in a situational and cultural contexts and serving certain purposes (Halliday & Hasan 1976)
• the psychological approach interested in cognitive processes of producing and comprehending texts, an approach being increasingly noticed by linguists (cf. Antos & Krings 1989; Strohner 1990; Krings & Antos 1992; Winter 1992)

(Hess-Lüttich & Niederhauser 1995: 349)

The first one is reminiscent of the pyramid of units in cultural semiotic "Theses" (1973). The second invokes Malinowski (the connection with Halliday is evident). The third is a bit more interesting, though psychological processes are notoriously slippery.

In the first chapter, Vater discusses various concepts of 'texts', starting with an everyday understanding of the term. Brinker (1992: 12), for instance, defines it as 'schriftlich fixierte sprachliche Einheit, die in der Regel mehr als einen Satz umfaßt' (p. 15). Vater, however, shows, by a number of illustrative examples, why this definition cannot meet modern scientific standards: there are texts in other than the written medium; the concept of a fixed linguistic unit focuses on the 'product' aspect neglecting the communicative 'function' aspect; texts as sequences of sentences exclude certain types of texts such as proverbs, etc. (Hess-Lüttich & Niederhauser 1995: 350)

In the case I'm interested in, "set phrases" are such fixed linguistic units but focus on them, e.g. in the guise of "phatic utterances", neglects the "flow of language", such as gossip, that is not "set".

This constellation of concepts ('Textwelt') is dependent on the knowledge and the intentions of the language users. (Hess-Lüttich & Niederhauser 1995: 350)

Neat... *click*

This approach cannot but define texts lacking these various structures as non-entities, as 'non-texts', whatever that may be. From a hermeneutic viewpoint, the critique is even more radical, of course, but that would be well beyond the perspective of a 'hard-core linguist'. (Hess-Lüttich & Niederhauser 1995: 351)

Made me think of the opposition of communion and communication turning the former into non-communication.

Instead, he returns to the solid ground of Prague School Linguistics and the analysis of theme/rheme-structures. He concludes by referring to [|] the approach of Klein and von Stutterheim (1992), who suggest extracting the theme of a text by answering an underlying question, explicit or implicit, the 'quaestio' of a text. But what about texts with a plurality of 'quaestiones', e.g., in everyday conversation, in literary dialogue or in phatic communication? Perhaps the authors would hesitate to regard texts of this sort as texts at all (cf. Hess-Lüttich 1981, 1984, 1985). (Hess-Lüttich & Niederhauser 1995: 351-352)

The authors correctly identify phatic communication and everyday conversation. It may be noted that phatic communion in literary dialogues has also been studied (cf. Urbanová 2007).

Taking a report in his daily newspaper, the Kölner Stadtanzeiger (a prominent source for many of his examples) as a starting point, Vater demonstrates how relations of co-reference and time reference work together in shaping something like a network of coherence. These networks remind us of the London School tradition which, however, is not referred to. Instead, Vater comes back to Klein and von Stutterheim (1991) and presents their concept of 'referential movement' as a further option to describe the gradual unfolding of information in texts. (Hess-Lüttich & Niederhauser 1995: 352)

A concept I did not know I needed, but I do. I think this "referential movement" can also be found in phatic communion, as when starting off with a commonplace about the weather leads the discussion to climate change and international politics. Malinowski's phatic communion, though, does not allow much referential movement, as the selfish speaker keeps yapping about whatever is on his mind; whereas in Ross's well-mannered type of conversation the topic flows, and speakers accommodate each other's interests.

Power, R. J. D.; Dal Martello, M. F. 1985. Methods of investigating conversation. Semiotica 53(1-3): 237-257.

This diversity is partly due to the book's origin: it is based on a conference, held at Bristol in 1979, with the broad title of 'Social Psychology and Language'. In part, however, the diversity of the collection is a deliberate reflection of the state of the discipline. Twenty years ago conversation was virtually virgin territory; since then, it has been invaded from several directions at once - from linguistics, sociology, psychology, philosophy, and artificial intelligence - each invader bringing its own characteristic repertory of empirical and theoretical techniques. (Power & Dal Martello 1985: 237)

Conversation was a "virtually virgin territory" in 1965? Difficult to believe.

Both in Conversation and Discourse and in the literature generally, two approaches seem to predominate: the linguistic approach, which derives in particular from Chomsky (1965) and Grice (1975), and the ethnomethodological approach, which was first articulated by Garfinkel (1967). Strictly speaking, however, ethnomethodology is an approach to the study of social life in general, not just conversation; its application to the analysis of conversation was originated by Sacks in the late 1960s, and has led to the development of the discipline that is now usually called 'conversational analysis'. (Power & Dal Martello 1985: 237)

Ethnomethodology is still appealing. And it looks like it's still making the rounds in certain quarters.

Speaking half seriously, we would summarize the situation thus: modern linguistics has tried its best to avoid becoming entangled in the complexities of conversation, but has been gradually forced in this direction by uncooperative data. The central feature of Chomskyan methodology (1965, chapter 1) is its restriction of the subject-matter of linguistics to the language-system itself; in effect, language is abstracted from the circumstances in which it is used. This methodological position was adopted for cogent reasons. Any given linguistic act results from the combined application of several systems of knowledge, of which the language-system is just one. (Power & Dal Martello 1985: 238)

Semiotics, too, doesn't want to deal with conversation all that much, but that is more understandable. The circumstances in which language is used = the context of situation. The combined application of several systems of knowledge reminiscent of Lotman's semiotic heterogeneity.

This method has produced rapid advances in syntactic theory, which has attained a high level of generality, precision, and mathematical sophistication. The development of a theory of meaning has proved more difficult. There are obvious respects in which the meaning of a sentence will vary from one context of utterance to another; moreover, 'meaning' is a vague word that can cover a number of different things. Because of these difficulties, linguists have not tried to solve the problem of meaning in one fell swoop; they have tried instead to isolate and make precise an aspect of meaning that is relatively independent of context. What they have done, specifically, is to focus on the descriptive aspect of meaning as opposed to the instrumental. Most utterances include, as part of their meaning, a descriptive component that corresponds to a possible state of affairs in the world; this component of meaning is sometimes called the 'propositional content' (Searle 1969). Utterances thus relate to the world in two ways. If Prince Charles says to Lady Diana 'Beethoven wrote five piano concertos', his utterance is related on the one hand to certain events that occurred between the years 1795 and 1809 in Vienna (or thereabouts), and on the other hand to an interaction taking place now, in 1985 (let us say) in Buckingham Palace. On this basis, a distinction is made between semantics and pragmatics: semantics deals with what an utterance depicts (in this case, Beethoven composing his five piano concertos), while pragmatics deals with what an utterance does (i.e. informing Lady Diana, among other things). Since the early 1970s semantic theory has attained considerable sophistication through the application of the formal techniques developed by logicians such as Montague (1974). (Power & Dal Martello 1985: 239)

Good overview. Will have to compare it with Morris' definitions of the semantic and the pragmatic. He also pointed out that the term "meaning" is a piece of putty and recommended the sharp arrows of "signification" and "denotation". Curiously, what Malinowski says about meaning in his definition of meaningless use of language is one of the most perplexing parts of it.

Many recent studies in pragmatics have adopted the idea that an important component of contexts is the set of propositions having what Grice (1967) calls 'common ground status'. As a conversation proceeds, the participants progressively augment a shared pool of information - the common ground - that can conveniently be regarded as a set of propositions; each successive utterance attempts to add something to this pool, and may also reflect in various ways the information that it already contains (e.g. in the choice between the definite and indefinite articles). (Power & Dal Martello 1985: 240)

This I've planned to approach with Morris' concept of "communization", i.e. what the communicants make common between them. Though, via Ruesch's expansion, it can include the life-experiences the communicants shared before the communication act.

This point can be most clearly expressed by using Quine's (1972) distinction between 'fitting' and 'guiding'. As Quine points out, a rule can be related to a phenomenon in two ways: it can fit it, as the motion of the planets fits Kepler's laws, or it can guide it, as the motion of the buses in London is guided (to some extent at least) by the timetable. What is special about social phenomena is that they not only fit certain general rules, but are also guided by them. Insofar as social life exhibits structure, it does so because the people who create it possess, either consciously or unconsciously, certain formal conceptions of social life that are characteristic of the culture. In a sense, all that the scientist investigating social interactions can do is to spell out what the ordinary person already unconsciously knows. (Power & Dal Martello 1985: 241)

Descriptive and prescriptive rules. That the scientist is merely trying to find out "what the ordinary person already unconsciously knows" is very evident in the study of nonverbal communication - no-one needs the rules of everyday interaction spelled out, and doing so may actually hinder normal interaction.

There do exist general rules that are characteristic of the social life of a culture, but these rules do not determine our behavior absolutely; consciously or unconsciously we may break the pattern. We cannot, however, escape the norms of our culture entirely, since even when we fail to conform to them, they remain the background against which our actions are interpreted. For instance, if someone greets you in a bar, the norm would be to return his greeting. This norm does not determine your behavior absolutely: you can if you wish sit still and make no response. But in doing so you have not escaped the influence of the norm; because it exists, you will be perceived not as having done nothing, but as having performed an action of great significance, an action from which inferences will be drawn about your feelings towards the person who greeted you. (Power & Dal Martello 1985: 241)

An illustration very pertinent to phatic communion.

By the simple method of tape-recarding natural conversations and analyzing them, they were able to identify features of social interactions which, although in a sense familiar, had never been explicitly noticed before. When one reads the early ethnomethodological papers on such topics as turn-taking (Sacks, Schegloff & Jefferson 1974), openings (Schegloff 1968), and closings (Schegloff & Sacks 1973), one experiences a revelation similar to that of the gentleman in Molière: we have been performing pre-closing sequences all our lives and not realized it. (Power & Dal Martello 1985: 42)

Establishing, prolonging, and terminating contact:

• Sacks, H.; Schegloff, E.; Jefferson, G. 1974. A simplest systematics for the organization of turn taking in conversation. Language 50(4): 696-735.
• Schegloff, E. 1968. Sequencing in conversational openings. American Anthropologist 70(4): 1075-1095.
Schegloff, E.; Sacks, H. 1973. Opening up closings. Semiotica 8(4): 289-327.

'Some strategies for sustaining conversation' (Wells, MacClure & Montgomery) is based on tape-recarded conversations between adults and small children. The aim of the study is to identify the strategies used by adults to keep such conversations going. These strategies, it is suggested, are grounded in certain structural features of ordinary conversation, and more generally in the phenomenon of 'prospectiveness' - that is, the degree to which an utterance looks forward or backward. (A question has a high degree of prospectiveness since it anticipates an answer; the answer instead is retrospective since it refers back to the question.) (Power & Dal Martello 1985: 244)

Jakobson (and Waugh's) interrogative and imperative orientations come to mind. Makes me thing if I should finally read The Sound Shape of Language from start to finish.

The second part of Owen's paper 'Conversational units and the use of well' (5) employs the techniques and concepts of conversational analysis in an attempt to identify the uses of the word 'well' when it occurs at the start of an utterance. The study was stimulated by the work of the linguist Lakoff (1973), who concluded that 'well' indicated some kind of insufficiency, either in the utterance itself or in the action to which it was a response. (Power & Dal Martello 1985: 245)

Jakobson's fingerprints are all over this one.

• Lakoff, R. 1973. Questionable answers and answerable questions. In: Kachru, B. et al. (eds.), Issues in Linguistics. Urbana: University of Illinois Press.

The main problem with this proposal is the vagueness of the terms 'move' and 'turn'. Owen takes the term 'move' from Goffman (1976: 272), whose definition is worth quoting in full, since it must be one of the vaguest in the history of science:

any full stretch of talk or of its substitutes which has a distinctive unitary bearing on some set or other of the circumstances in which the participants find themselves.

Owen does improve on this to some extent with her proposal that a move is either the first or second part of an dajacency pair. So far as we can see, what Owen and Goffman intend is that a move should be a minimally tangible contribution to a conversation; the problem is to elucidate what 'tangible' means here. If a person says 'I entirely -' and is then interrupted, he has made no impact upon the development of the topic; he might as well have saved his breath. If instead he says 'I entirely agree', he has done something which is likely to register in the long-term memories of the other participants and to influence the future course of the conversation. Can this intuitive distinction be made precise? It might turn out that one can do no better than distinguish degrees of tangibility. (Power & Dal Martello 1985: 249)

A possible criterion to examine phatic communion from a novel angle. By definition, it should be completely intangible, the topic so irrelevant that it doesn't "register in the long-term memories of the other participants". Whether a move influences the future course of the conversation - the "referential movement", so to say - is a question distinct from this.

This view seems to be shared by Owen (5), who suggests that the syntactic-prosodic account of transition relevance places given by Sacks et al. (1974) will have to give way to 'some kind of functional explanation' (p. 100). (For those unversed in ethnomethodological terminology, what this means is that the point where the speaking turn can pass naturally to another participant is determined by functional features of the utterance-in-progress rather than determined by syntactic or prosodic features. 'Functional' here is synonymous with 'pragmatic'.) (Power & Dal Martello 1985: 250)

Jakobson's "contoural feature" is prosodic. What does "pragmatic" mean here, then?

A 'functional' explanation will presumably be one that refers to what utterances do. This characterization is of course too vague: as Austin (1962) showed, one can describe 'what utterances do' on a number of different levels. For instance, a boy uttering the word 'Mum!' may be simultaneously articulating a phonetic form, articulating a word, referring to his mother, and summoning his mother. What concerns us here is clearly the higher-level social goals of utterances (e.g. summoning) rather than the 'rhetic', 'phatic', and 'phonetic' acts (to use Austin's terminology) by which these goals are realized. Utterances change the social or interpersonal state of affairs between speaker and hearer: more briefly, they change the social context. A precise functional characterization of an utterance therefore requires a conceptual apparatus for describing social context. It is through lack of such an apparatus that the theories described above had to resort to vague intuitive concepts like 'move' and 'GIVE'. (Power & Dal Martello 1985: 250)

It sounds as if it is this sense that Malinowski calls his invention (purely) "social". It almost seems to beg the question: why is the boy summoning his mother?

Two of the papers in Conversation and Discourse (Werth (7) and Wilson & Sperber (8)) are concerned with elucidating the concept of relevance in conversation. Wilson and Sperber (whom we shall focus on here) suggest that the relevance of an utterance is determined, roughly speaking, by the number of pragmatic implications it has. If this idea seems arbitrary, consider that when somebody says something that has no apparent relevance, a common response is 'So what?', which is a brief way of saying 'What does that imply?' (Power & Dal Martello 1985: 251)

So this is where relevance theory came to the fore.

Sacks et al. suggest in a footnote that intonation is also important, but as Owen (5) remarks, what seems to be needed is a functional explanation. It is here that Wilson and Sperber's work on relevance might prove useful. It seems quite likely that the duration of the speaking turn is governed by the requirement that the speaker must be permitted to establish the relevance of what he is saying. In terms of Wilson and Sperber's definition, this would mean that transition-relevance places would occur at points where the hearer was able to derive pragmatic implications from the semantic content that had so far been presented. (Power & Dal Martello 1985: 252)

Are they implying that conversation is not a sequence of observations about irrelevant happenings?

The agenda concludes with the following puzzling question, which we quote verbatim: 'Can the several instrumentalities be mapped into a bounded multi-dimensional space?' Is Grimshaw serious? Or are we being Garfinkelled? The only attempt that is made to narrow down the field of investigation is the emphasis upon utterances as 'instrumentalities' - that is, as attempts to manipulate the hearer. Even here, unfortunately, it is unclear what has been achieved, since presumably all utterances in a conversation are designed to have some effect on the hearer. (Power & Dal Martello 1985: 254)

What effect is intended, though? Malinowski's preoccupation was the intent to impress. Yet according to the comparison with Dewey's functions, phatic communion should ideally not intend to manipulate the hearer.

Our final criticism concerns the introduction of an ew and eccentric terminology for the concepts of speaker, hearer, and utterance: the terms proposed are respectively SOURCE, GOAL, and INSTRUMENTALITY. Apart from the weird choice of GOAL as a term denoting the hearer, the trouble with this proposal is that it achieves nothing. One does not advance science by replacing imprecise words that everyone understands with equally imprecise words that hardly anyone understands. (Power & Dal Martello 1985: 154)

Oh boyy. Phatic communion vs small talk.

Makkai, Adam 1987. Idiomaticity and phraseology in post-Chomskian linguistics: The coming-of-age of semantics beyond the sentence. Semiotica 64(1/2): 171-187.

What happened is actually disappointingly simple: grammarians came to realize that actual human beings in real societies communicate with one another because they have some socially relevant message to transmit, in other words, that humans use language as a semiotic device in order to acquire goods and services; to take inventory of their environments; to engage in phatic communion, etc. (See Halliday 1973, 1974, and 1975) (Makkai 1987: 171)

Disappointing indeed. Phatic communion also turns out to be a means to an end (admiration of one's fellows) rather than an end in itself, which Malinowski is careful to qualify with "almost".

• Halliday, M. A. K. 1973. Explorations in the Function of Language. London: Edward Arnold.
• Halliday, M. A. K. 1974. Learning How to Mean. London: Edward Arnold.
• Halliday, M. A. K. 1975. Language as Social Semiotic. London: Edward Arnold.

Whenever a 'Context of Situation' became necessary in order to explain why a sentence is grammatical or why two sentences are coherent, the grammarian, still based on the rationalistic premises of Chomskian thought, had no alternative but to conclude that there was something 'outside language' that somehow had a determining effect on whether the language used at the given moment was sensible or not. A tremendous variety of expressions were coined ranging from 'felicity condition', to 'semantic grammaticality', to 'social acceptability'; what most scholars agreed on was that 'real life' had the upper hand over grammatical speculation in the abstract. (Makkai 1987: 172)

Eerily reminiscent of Jakobson's case with the emotive function - emotions are, after all, "extralinguistic".

Consider a simple example. A meets B in the street at 11 p.m., going home from the office where both have worked late. A says to B good night. B looks at A strangely and says hi there, and good night to you, too. (He also thinks to himself 'A will never learn English.' A's words good night are perfectly good English: they constitute an act of 'salutancy', and the time is nigh. What went wrong? A doesn't seem to know that good night in English is only appropriate on taking one's leave, or on departure. Thus, no matter how late it is, you first say hello, there!, hi!, what do you say Joe? - or anything customary; good night, along with so long!, nice seeing you! see you later! etc., is only said upon separation of the two speakers. (Makkai 1987: 172)

Another pertinent illustration. This one actually supports Malinowski's implicit thesis that contexts of situation are connected with typical utterances. That is what makes an utterance appropriate - the context.

A has a problem. Was this a bit of English grammar? Was this a bit of English semantics? What were these restrictions all about? Maybe it's an English-specific problem, he thinks, and tries to find something of a comparable nature in Spanish. So he tries buenos días!, buenas tardes! and buenas noches! at the expected, appropriate times (morning, day - early afternoon, and evening and night, respectively) and also at inappropriate times. Thus one morning at 6 a.m. before sunrise he greets a Hispanic colleague by saying buenas noches! The colleague smirks and answers 'Oh, you've been up all night and are just going to sleep?' from which A draws the conclusion that dias, tardes, and noches, 'daytime', 'late hours' and 'night-time' are relative to the speaker's obvious or expected awareness of what time of the day it actually is. Having or not having a watch and knowing what time it is - although normally expected of normal people in normal circumstances - cannot be a grammatical skill. If someone has been kept in a jail cell where there is no window with a 100 Watt burner on all the time, so that the person falls asleep off and on not knowing even when a new day started, upon emerging from the cell he cannot be held responsible for selecting días, tardes, or noches correctly. (Makkai 1987: 173)

Something like this I've experienced in night-shifts. I'll say "Morning!" because I've just woken up and arrived at work in the late evening, but objectively it is not the morning, and neither is it a subjective "morning" to the day-shift workers shortly going home.

In a sense, then, there was no real need for inventing 'pragmatics', because it was understood throughout the 19th century in philological linguistics throughout Europe. In the work of Firth as well as in that of his followers, such as Mitchell and others, the Malinowskian notion of the 'Context of Situation' was fully employed and consciously practiced in 1930. The intervening half a century to 1980, then, is a period of retrogression in terms of a realistic view of human language rather than five decades of heroic breakthroughs. The attempt to formalize human language in the meta-language of symbolic logic and set theory came at a high price: We were encouraged to forget that our ancestors had solved many problems for us and we thus found ourselves in the ridiculous position of having to re-invent the wheel, in order to account for the simplest data of natural language. (Makkai 1987: 173)

That's what I like to hear! Exactly what I was thinking after reading the previous paper and before reading this one. Malinowski and Jakobson and the intervening decades managed many breakthroughs that appeared revolutionary in the 1960s and even later. "Pragmatics" is a special case because it can be argued that it saw a case of "simultaneous emergence" in Malinowski. His unique pragmatism, on the other hand, is not very clear nor very well understood.

It is also evident that these theories can, if one has the patience, be mapped onto one another. (Makkai 1987: 174)

More or less my currently set goal: to show that some interesting problems that "broke through" later on (e.g. common ground/communization) were proposed or even solved some time earlier. My task, as it were, is to map "modern" theories onto ones current a century ago.

What is pragmatics, really? Pragmatics is the recognition of the fact that grammar alone does not enable a spaker of a language to make good sense in that language, to be effective, sufficiently manipulative, funny, entertaining, in short what is expected of a successful native speaker. (Makkai 1987: 174)

Linguistic wit as a qualifier of nativeness.

Many linguists of the younger generation, when confronted with the untameable jungle of the lexicon, in order to be 'elegant' and not fall victims to criticism that they are interested in 'random facts', escape into the prestige-jargon created by Searle and Austin and other post-Chomskian logicians and refer to language facts that the theory has no way of handling as 'pragmatics'. (Makkai 1987: 175)

"Prestige-jargon" a very well-deserved designation. The same could be said of the jargon of the ethnomethodologist.

It is no wonder in view of the foregoing that phrases and idioms should be the prime hunting ground for the 'pragmatist'. What else could be done with idioms and phrases than relegate them to the obscure area of 'special know-how'; the native speaker's awareness that 'President kicks bucket' is not an acceptable newspaper headline, whereas 'President slain' or 'President assassinated' both are. (Makkai 1987: 175)

Set phrases do do them tricks.

[...] and the description must contrast 'table' with desk for English, since many European languages use 'table' for desk as well with some prefixed material or other, e.g., German Schreibtisch, Hungarian író, 'for writing', and asztal, 'table', respectively. It is an accident of English that writing-table did not emerge and is substituted for by desk; similarly the fact that my dining table is not referrable to as my eating desk. Note, incidentally, that these forms are not 'ungrammatical' or incomprehensible; the effect they give is that you are listening to a struggling foreigner who tries to translate into English what he might have said in his language. (Makkai 1987: 175)

Very much like my recent mullings over how English has, by way of accident, "newspapers" and not "timewritings".

Another way of expressing the same notion would be to say that the meanings of a lexeme form a complex network of relationships. Such a network, by the way, need not necessarily be drawn; it is sometimes sufficient to give a verbal exposition; with modern minicomputers all sorts of other arrangements are also possible (see Makkai 1980). (Makkai 1987: 176)

Related to the "constellations of concepts". At some point, perhaps in my BA thesis but perhaps elsewhere, I'll have to draw out the etymological hypotheses of "phatic" that I've postulated (not sure if there are 4 or 5 of them at this point, simply "suffering" a skinflints' boasts being the fifth and perhaps even the simplest).

Many of these forms involve - on the surface - simple and familiar grammatical processes (or relationships; as you wish), e.g., adj. + N. Some will have unusual stress patterns such as White = House, black = bird and Yellow = Stone as opposed to the 'literal' white + house, black + bird, and yellow + sto, but this difference in stress is not a predictable concomitant of the semantic opacity or transparency of a given compound. This is well known material and is adequately treated in the chapter on idiomaticity in Hocket (1958). The ubiquitous 'arbitrariness of the sign' haunts the would-be-regularizer of natural language at every step of the way, because erratic spellings, incomplete patterns, and reversals of expected accent patterns all add to the aggravation caused by logically generatable but actually nonexistent forms. (Makkai 1987: 176)

Still on the meta-level, the "erratic spellings, incomplete patterns" etc. call to mind the immense variability of nouns "phatic" is attached to: communication, exchange, utterance, interaction, speech, etc. come to mind. How does this come about? One person reads Malinowski, mis-remembers the exact term, exchanges the noun for another, and the next person reads that and thinks that's it?

Certain set phrases in English whose meanings are not predictable automatically and are also subject to erroneous decoding, are simultaneously 'parts of speech' of a complex nature, that is, they are expanded verbs, nouns, adjective-like nominoid constructions modifying other nouns. A typical example is come down with as in grandma came down with the flu. I called this a lexemic idiom (Makkai 1972) for the following reasons:

It is substitutable with a single transitive verb with no loss of denotative or 'legal meaning' as in grandma contracted the flu.
1. It is not alone in English; another related type is go through which means 'persevere' as in the general went through with his plans.
2. The pattern includes other three-part 'verbs' with other formatives such as for, e.g., go in for, as in John goes in for stamp collecting meaning 'John cultivates a stamp collection'.

From these examples one establishes the lexicographic equivalences

• come down with = 'contract'
• go through with = 'persevere with regard to x'
• go in for = 'cultivate'.

(Makkai 1987: 179)

Set phrases! Of course one could substitute "Good morning," "Good day," and "Good evening" with a simple "Greetings!"

Hudson argues that this identity in parsing 'proves' that come down with can only be a sememic idiom, since come, down, and with all behave perfectly regularly and literally. It has been suggested by others that the idiomaticity of these expressions does not reside so much in the come down with part as in the succedent object suitcase (literal) and flu (idiomatic). I can now see why so much misunderstanding has arisen in connection with these idioms. Their behavior is indeed strange, to say the least. It seems that one can come down with something quite literally as long as it is an object, but that coming down with something changes meaning figuratively and becomes 'contract' or something similary, if the semantics of the direct object noun has some such feature as 'communicable disease of non-serious nature and not too long a duration'. (It is interesting to remark that you cannot come down with congenital heart disease while you can come down with the flu, the measles, rubeola, etc.) (Makkai 1987: 179)

One doesn't simply "come down" with COVID-19. Or, as the idiot president's idiot sun recently put his own infection, "caught the 'rona'.

Let us take a look at an example. The form is O-genki desu-ka? Its communicative function is a greeting; the speaker initiates the exchange. Its communicative effect is that the addressee must answer; usually with the formula o-kagesama de. Outer fixedness is omitted. The form's inner fixedness is very strong: the honorific prefix o has to be present and the form is inflected. The literal meaning is 'are you healthy' or 'the health, is it there?'. Abstract use is left blank; under usage conditions we see 'greeting among people who know one another: offers the possibility to the opening up of conversation, no explicit answer regarding the addresse's health is expected.' (Makkai 1987: 184)

"Greeting" is a function?

This, then, works like how are you? in English. Casual acquaintances will ask one at the bus stop or in an elevator, and to answer by reciting one's medical condition would be a major faux pas; one answers fine, how are you? and likewise does not expect an answer. Malinowski called these exchanges phatic communion. It would be most interesting to see how such forms compare on a cross-cultural, multilingual basis. (Makkai 1987: 184)

Even includes the classical example, that replying to "How are you?" with how you actually are is not appropriate.

The introduction is lucid, fair, thorough and quite readable and gives explicit explanations as to why people use routine formulae in their every day speech. Coulmas feels, as many of us in idiomaticity/phraseology do, that real creativity is not so much in the application of syntactic rules to randomly mastered lexis but rather the other way round: lexis grows and is created by writers, individuals, and social situations, while the syntax grudgingly follows along. Coulmas is a skillful user of the Malinowskian notion of 'the context of the situation'. As a mumber of our post-1956 generation, he, too, sees 'pragmatics' where there used to be plain common sense in the olden days. (Makkai 1987: 185)

Noteworthy because there are a lot of writers who do not use the concept of the "context of situation" with full knowledge of what they're talking about.

It seems likely that as interest in syntax gradually lessens, scholars will pay more and more attention to metaphors, fixed formulas, etc. One of the great giants of human thought, hithero totally neglected in American linguistics, is the 18th-century philosopher, Giambattista vico, who in his La scienza nouova (1968) suggests that human language started with wordless gestures and hieroglyphic writings. This state of affairs would have corresponded to the great theocracies such as ancient India and ancient Egypt. This, in Vico's view, was followed by the heroic age (Homer, etc.), which gave rise to speaking in metaphors and similies ('rosy-fingered dawn', 'ox-eyed Hera', etc.). The heroic age was followed by democracy which also created Roman Law and brught in alphabet scripts and logically reasoned grammars via mankind's ability to think abstractly. The democratic age is followed by chaos, or confusion, from which a new cycle starts. (Makkai 1987: 185)

"Fixed formulae" a synonym of "set phrases". Vico sounds more and more interesting. I've noticed recent studies of him.

If natural languages on the planet are sign systems, these sign systems have internal subsystems that deal with complex signs that somehow differ from the rest inasmuch as they are grammatically fixed, are surprisingly hard to make up, are tricky to decode, or often impossible to translate, and are greatly characteristic and typical of the langage in whose major sign-system they occupy a significant section. (Makkai 1987: 185)

"Set phrases" in the cultural semiotic ecosystem of language. The Trobriand "Whence comest thou?" comes to mind.

Constituent words of idioms will be rapidly located in environments other than the idiom in which they are contained and thus the reinvestment ratio of single words and/or morphemes will be measurable in terms of frequency in various registers, for example, newspapers, novels, taped conversations at parties, street gangs, etc. (Makkai 1987: 186)

"Are words in Phatic Communion used primarily to convey meaning, the meaning which is symbolically theirs?" (PC 6.3) - If they occur outside of phatic communion, too, then perhaps.