HENRY SWEET

HENRY SWEET AND AMERICAN STRUCTURALISM: A BRIEF COMPARISON.

Luis Quereda Rodríguez-Navarro

Granada University

                If there is one word which best summarizes Sweet's work, this word is pioneer. This is, at least, what Bolton and Crystal believe. They claim:

Sweet was a pioneer student of the English language who did much to further knowledge of both its history and its living structure. (Bolton & Crystal, 1969:8).

                  Wrenn seems to be of the same opinion. Talking about his life and work, he says that Sweet was:  

One of the [Philological] Society's greatest pioneer leaders and also, in my opinion, the greatest philologist that our country has so far produced. (Wrenn, 1967:150).

                  According to definition no.3 in The Oxford English Dictionary a pioneer is "one who goes before to prepare or open up the way for others to follow; one who begins, or takes part in beginning, some enterprise, course of action, etc.; an original investigator, explorer, or worker, in any department of knowledge or activity." Therefore, the first thing that we shall have to do, following this definition, in order to appreciate to what extent Sweet's work was original and innovative will be to evaluate the work of other grammarians of his time. Since Sweet's work is pioneer in more than one field, we shall concentrate in this paper on descriptive linguistics and, more specifically, on the description of English as a living language.

              The grammatical work which was performed in Britain during the nineteenth century was not very innovative. Michael sums up the situation as follows:

    Most grammars of English published in Britain during the 19th century are dull ... There were a great many grammars, issued in very large numbers. They were repetitive; many were merely commercial ventures, scholastically naive. (Michael, 1991:11).

              Michael (1991:12) has catalogued 856 different grammars, but in his view the grammars published during the nineteenth century must have been at least about one thousand.^[1] Despite the vast number of grammars, their contribution to the analysis of the English language was very poor. In this respect, Michael concludes:  

The vast number of grammars contrasts with the uniformity of their contents. Of all the subjects in the school curriculum English grammar was the most rigid and nchanging ... Teachers had insisted, for two centuries, on writing grammars which added little or nothing to what had gone before. (Michael, 1991:13).  

            Work produced abroad was different, but did not constitute an example to be followed. The best English grammar of the nineteenth century was Maetzner's Englische Grammatik (1860-65), which was translated into English in 1874; in its preface, the translator, Grece, praises the work done in Germany on the English language. And Sweet, who was not generally enthusiastic about German scholarship (1891:viii-ix), acknowledges his debt to Maetzner in the preface to his New English grammar (1891:xiii). But Sweet nevertheless insisted that a description of present-day English, and not "antiquarian philology", was his aim (1891:207). Maetzner's grammar was not the first and only one written within the comparative-historical paradigm,^[2] but it was by any standard the most comprehensive treatment of English in the century. However, as Leitner (1986) explains, Maetzner's grammar had main three drawbacks. The first was due to its historical orientation: "Maetzner treats countless constructions, forms and usages that would be better considered archaic, or even non-existent, from a more narrow temporal perspective." (Leitner, 1986:416). The second was that "his categories were defined in notional, not formal, terms," (Leitner, 1986:419), and the third is "that his treatment of pronunciation was based on letters." (Leitner, 1986:419).  

            Taking all this into account, we can conclude that Sweet's position is original and innovative when compared to the work which had been done before him, not only in Britain but also elsewhere. But, as the Oxford Dictionary definition tells, this is not enough to be considered a pioneer. To be a pioneer, one has to initiate in some way what others will continue in the future. And this is precisely what Sweet did. In order to show this, we shall compare Sweet's reaction to traditional school grammar with that of American structuralists.

            American structural linguistics has normally been presented as the revolution linguistics needed to get rid of all the obsolete ideas traditional grammar represented. In this respect, Gleason says of Fries' work:  

Fries' American English Grammar was an important event. It opened up a whole direction of development. (Gleason, 1965:19).  

The American English Grammar, however, does not give a system complete enough to provide the basis for the new curriculum he envisioned. To do so, it would have to have included a description of sentence structure ... It was twelve years before the continuation appears as The Structure of English ... The delay was probably fortunate. It was a far better book in 1952 than it could have been earlier. Before 1940 American descriptive linguistics had very little to contribute to the analysis or statement of syntax. The decade before The Structure of English appeared was one of intense activity in this field. Fries was able to make use of some of the newly developed techniques. (Gleason, 1965:20).  

The emphasis in The Structure of English is clearly on sentence structure. For this reason the whole system is best known as 'structural grammar'. Looked at from the point of view of school grammar, it was a new and radical innovation. Hence it became known among English teachers and school administrators as the 'new grammar'. Because of Fries' insistence on the principles of linguistics ... the scheme also came to be identified as 'linguistics'. (Gleason, 1965:80).  

In contrast, Gleason is not so enthusiastic when he talks about the scholarly tradition, a movement within which Sweet is normally classified: 

  Much of the work of the scholarly traditional grammarians has suffered from the lack of an adequate framework within which they might assess the relevance of the examples studied. Only an adequate basic outline continually examined in the light of a general theory of language could provide this. But few gave any attention to such questions. The general structure was accepted from past authors without as much critical consideration as it deserved. The focus of attention for the European grammarians has always been strongly on the details, particularly the uncertain or controversial points that marked the boundaries of current knowledge. (Gleason, 1965:77).

This is, unfortunately, the general picture that has normally been presented to us. We believe, however, that this picture has been somewhat exaggerated, and that it is definitely not the whole picture, since it ignores the European reaction against traditional school grammar, which began precisely with Sweet more than fifty years before the American structuralist movement started, as we shall try to prove with this quick comparison between the tenets held in Sweet's work and those in American structural linguistics.  

            American structural linguistics is usually presented as a reaction against traditional grammar. For structuralist linguists, traditional grammars were wrong in their principles and in their methods. Traditional grammarians were wrong in their whole approach to language, in their notion of what a language is and how it may be adequately described. As a reaction to this, American structuralists made claims for a scientific approach to language. Language should be studied to know and to understand how language works. There is a desire to make the study of language both scientific and autonomous. But this desire can already be seen clearly in Sweet's work. His introductory words to New English grammar leave no doubt in this respect:  

This work is intended to supply the want of a scientific English grammar, founded on an independent critical survey of the latest results of linguistic investigation as far as they bear, directly or indirectly, on the English language. (Sweet, 1891:v).  

            Let us now examine in a more detailed way whether the concept that Sweet had of a scientific grammar was very different from that of American structuralist grammarians.  

            American structuralists thought that traditional school grammars were unscientific because they were typically normative and gave much of their space to correcting errors. As a reaction, structuralists thought that there could be no correctness apart from usage. For them, language should be described as it is spoken, and never as some grammarians might think it "ought" to be spoken. Fries' position is self-explanatory:  

The point of view in this discussion is descriptive, not normative or legislative. The reader will find here, not how certain teachers or textbook writers or "authorities" think native speakers ought to use the language, but how certain native speakers actually do use it in natural, practical conversations carrying on the various activities of a community. (Fries, 1957:3). 

But Sweet's opinion does not differ much from Fries': 

As my exposition claims to be scientific, I confine myself to the statement and explanation of facts, without attempting to settle the relative correctness of divergent usages. (Sweet, 1891:xi).  

In considering the use of grammar as a corrective of what are called 'ungrammatical' expressions, it must be borne in mind that the rules of grammar have no value except as statements of facts: whatever is in general use in a language is for that very reason grammatically correct. (Sweet, 1891:5).  

                Since American structuralists think that grammar has to do with the way English is spoken, and not with how it 'should' be spoken, differences in language practice should be accepted and different registers should be established. Fries (1940), for example, distinguishes at least four different registers: historical, regional, literary and colloquial. And again this is simply what Sweet (1891:201-203) does when he distinguishes first between languages and dialects, secondly between standard, refined and vulgar speech, and finally between literary and spoken language.  

            American structuralists also thought that traditional school grammars were unscientific because they mixed up synchronic with diachronic information, often giving predominance to diachronic forms. As a reaction, structuralist linguists clearly differentiated synchronic description of languages from diachronic studies. For them, it was important to define the stages of description and to distinguish between descriptive and historical linguistics. Descriptive linguistics is interested in how people speak now and not in how people spoke in past stages of the language. Fries claims:  

[The study here presented] assumes that the only method to attain really good English ... demands constant observation of the actual practice of the users of the language. (Fries, 1940:24). (Our emphasis).  

Sweet's New English grammar, although historical, is not, as he himself qualifies it:

One-sidedly or fanatically historical. The old belief in the value of historical and comparative philology as an aid to practical study of languages has been rudely shaken of late years. (1891:viii).  

            Therefore, to be consistent with this, Sweet takes considerable pains to keep descriptive apart from historical grammar. Thus, in the first volume of his New English grammar he first deals with those aspects related to the descriptive grammar of English (pp. 1-175), leaving the second part for historical considerations. It is only in his second volume that he studies the syntactic behaviour of the different periods of English together. Nevertheless he always tries not to mix them up. Sweet's plans are clearly stated in the following quotations:  

It is evident that all study of grammar must begin with being purely descriptive. Thus it is no use attempting to study the history of inflections in different periods of a language or in a group of cognate languages, if we have not previously got a clear idea of what inflections really are; and it is neither profitable nor interesting to compare languages or periods of languages of which we have no practical descriptive knowledge. (Sweet, 1891:204).  

In studying grammar it is important to keep the descriptive and the historical view apart. The first object in studying grammar is to learn to observe linguistic facts as they are, not as they ought to be, or as they were in an earlier stage in the language. When the historical view of language gets the upper hand, it is apt to degenerate into one-sided antiquarian philology, which regards living languages merely as stepping-stones to earlier periods. (Sweet, 1891:207).

                Therefore, Sweet's position does not differ in theory from that of American structuralist linguists. It only differs in practice, since he is interested in the description of the different historical stages of English, whereas American structuralists are not.

              American structuralists also thought that traditional school grammars were unscientific because they transferred the categories of Latin to such a different language as English. As a reaction, they assumed that every language must be described in and by itself. For them, there are no universal categories. The concepts "noun" or "adjective" in English must be different from those in French, since their real value does not lie in themselves but in their specific position within the system. Every language should be considered as a system of relations. Every unit, every element in this system has no value by itself, if isolated. Its meaning has to be established in relation to all the other elements in the language. In this respect, Fries claims:  

Another of the basic assumptions of our approach ... is that formal signals of structural meanings operate in a system - that is, that the items of form and arrangement have signalling significance only as they are parts of patterns in a structural whole. (Fries, 1957:59-60).  

            Sweet may not go as far as structuralists in this, but his position clearly seems closer to structuralists than to traditional school grammars: 

I may state at once that I consider the conventional treatment of English to be both unscientific and unpractical, starting as it does with the assumption that English is an inflexional language like Latin or Greek ... It was assumed, for instance, that as Latin had five cases, English must necessarily have just as many and no more ... (Sweet, 1876:19).  

Every language has the right to be regarded as an actual, existing organism ... The only rational principle is to look at the language as it is now, and ask ourselves, How does this language express the relations of its words to one another? (Sweet, 1876:20).  

                American structuralists considered that traditional school grammars were unscientific because they based their descriptions on obsolete literary written texts. As a reaction, American structuralist linguists based their descriptions on objectively observable data, paying special attention to current speech. In order to avoid the dangers implicit in traditional grammar, American structuralists had as their aim the description of the current spoken language of an individual or of a community.^[3]  

 In this respect, Fries (1940:27) considers that "the ideal material, of course, for any survey of the inflections and syntax of Present-day American English would be mechanical records of the spontaneous, unstudied speech of a large number of carefully chosen subjects." Fries (1952) is based on recorded telephone conversations. Sledd states this tenet of structural grammar in the following way:  

Men talked long before they wrote; and ideally, if we want to understand the tools of communication and expression which our writing system gives us, we should first understand the resources of our speech, from which our writing is derived. (Sledd, 1959:20).  

            It is the spoken language, then, that has to be analysed. But, as we have emphasized, both Fries and Sledd used the word ideal. This means that, although no one doubts that it is the spoken language that should be described, it is always easier to base our description on written language. Fries (1940) justifies his corpus of examples by saying:  

The use of any kind of written material for the purpose of investigating the living language is always a compromise, but at present an unavoidable one and the problem becomes one of finding the best type of written specimens for the purpose in hand. (Fries, 1940:27).  

Even nowadays, more than one hundred years after the publication of Sweet's New English grammar, grammarians are basically working with written data. Most machine-readable corpora available for research in Present-day English have data from written sources.^[4]  

            It is, then, obvious that the possibilities Sweet had of basing his investigations on spoken language were very scarce. But this does not mean that Sweet did not realize the significance of the spoken language in descriptive linguistics. He is clear in this respect: 

The study of a language should always be based - as far as possible - on the spoken language of the period which is being dealt with. (Sweet, 1891:203).  

It is well known that the German grammars make a complete confusion between the different periods of Modern English, all grammars ... ignoring the distinction between the literary and spoken language. This again has been completely reformed in the present grammar, in which the spoken language has had its proper importance assigned to it. (Sweet 1891:x).  

Due to the primacy of written language, traditional school grammar continues to neglect phonology, the basic study of the sounds of speech, and sometimes to confuse speech with writing. As a reaction, American structuralists considered phonology as the starting point of any investigation. Bloomfield (1933:162) claims that "linguistic study must always start from the phonetic form and not from the meaning." And probably for this reason phonology is the field where structuralists made more advances. According to Gleason (1965:40) it was the phoneme principle that gave the first workable basis on which to build a modern theory of descriptive linguistics. 

            However, Sweet, like structuralists, begins with the phonetic level in his article Words, logic and grammar (1876) and shows, among other things, the independence of the phonetic and logical systems in language. Sweet recognized the importance of the spoken language and of the study of phonetics, pointing out the role that stress and intonation have in carrying many differences in meaning. In his New English grammar, Sweet states:  

An essential feature of this grammar is that it is on a phonetic basis. It is now generally recognized, except in hopelessly obscurantist circles, that phonology is the indispensable foundation of all linguistic study. (Sweet, 1891:xii). 

The first requisite is a knowledge of phonetics, or the form of language. We must learn to regard language solely as consisting of groups of sounds, independently of the written symbols. (Sweet, 1876:9). 

            Sweet's interest in the spoken language is reflected in his grammar in the several sections which he devotes to stress and intonation of words, word groups and sentences. Sweet placed great emphasis on the analysis of sounds used by speakers of English to communicate their thoughts. He realized that commonly used English spelling was an unacceptable means of representing the speech sounds made by speakers of English, and he consequently searched for a way to establish a system of representing speech sounds with the greatest possible fidelity to their spoken form. The key requirement for such a system was that each symbol represented one and only one sound. The result was Sweet's phonetic alphabet. As has already been mentioned, Gleason thinks that it was the idea of the phoneme that made structuralist grammarians so successful. Nevertheless, according to Wrenn, Sweet may virtually be regarded as a co-equal with Baudouin de Courtenay in discovering the phoneme. Wrenn affirms: 

One cannot be sure whether Sweet or de Courtenay was the first to realise this new and most important conception, since they worked ... in entire ignorance of each other's studies. But in 1877 Sweet had clearly recognised ... the idea which is at the base of de Courtenay's elaborately reasoned explanation of the phoneme, and 1877 was the year when probably both scholars definitely reached their findings - Sweet in print only by implication and as a purely practical aid, and de Courtenay theoretically and at length, though only in then unpublished lectures. (Wrenn, 1967:159).  

            But the most important objection American structuralists had against traditional school grammar was that traditional analyses were vague, full of notional definitions and lacked grammatical information. Traditional grammar used different criteria to classify words (semantic for verbs and nouns and syntactic for pronouns, adjectives or adverbs). In order to avoid the dangers implicit in traditional grammar, American structuralists limited the area of language to be described by emphasizing language form as the single, objective, observable and verifiable aspect of language, thus relegating meaning to a subordinate place. For them, linguistic analysis should begin with an objective description of the forms of language and move from form to meaning. A good description of a language must be a description of its forms: 

One of the basic assumptions of our approach here to the grammatical analysis of sentences is that all the structural signals in English are strictly formal matters that can be described in physical terms of form, correlations of these forms, and arrangements of order. (Fries, 1957:58).  

            In general terms structuralists rejected meaning, arguing that what linguists cannot see or measure, as is the case with notional meaning, cannot be investigated, and that a scientific description of meaning is impossible not just in practice but also in principle. However, they realized that meaning is an essential part of language which cannot be discarded so easily. Fries states his position in the following terms: 

This challenge of the conventional use of meaning as the basic tool of analysis must not lead to the conclusion that I have ignored meaning as such, nor that I deny that the chief business of language is to communicate meanings of various kinds, and that the linguistic student must constantly deal with meanings. (Fries, 1957:8). 

Bloomfield is perhaps clearer than Fries: 

To put it briefly, in human speech, different sounds have different meanings. To study this co-ordination of certain meaning is to study language. (Bloomfield, 1933:27).  

Sweet defends the same kind of statement with very similar words: 

Language and grammar are concerned not with form and meaning separately, but with the connections between them, these being the real phenomena of language. (Sweet, 1891:7). 

This idea is later developed when Sweet claims: 

The business of grammar is to state and explain those relations between forms and meanings which can be brought under general rules. Theoretically speaking, these two - form and meaning - are inseparable, and in a perfect language they would be so; but in languages as they actually are, form is never in complete harmony with meaning - there is always a divergence between the two. This divergence makes it not only possible, but desirable, to treat form and meaning separately - at least, to some extent. (Sweet, 1891:204).  

            No one can doubt that Sweet's approach to grammar is formal. Sweet, for example, considers parts of speech both formally and functionally. He says: 

As regards their function in the sentence, words fall under certain classes called parts of speech, all the members of each of these classes having certain formal characteristics in common which distinguish them from the members of the other classes.... [parts of speech] have inflection of their own distinct from those of the other parts of speech (I grow, he grows, grown); ... each part of speech has special form-words associated with it (a tree, the tree; to grow, is growing, has grown); and ... each part of speech has a more or less definite position in the sentence with regard to other parts of speech. (Sweet, 1891:35).  

For Sweet, form is normally the only valid criterion for classification. Sweet's work is full of examples of this attitude. Let's analyse two of them. In the first place, for Sweet, the distinction between transitive and intransitive verbs is purely formal: 

When an intransitive verb requires a noun-word to complete its meaning, the noun word is joined to it by a preposition, forming a prepositional complement, as in he came to London; he looked at the house, I thought of that ... We can see that the distinction between transitive and intransitive is mainly formal, for think of and the transitive verb consider in I consider that have practically the same meaning, and think itself is used transitively in some phrases, as I thought as much. So the slight difference in meaning between he looked at the house and he saw the house has nothing to with one verb being intransitive, the other transitive. (Sweet, 1891:91).  

And, in the second place, it is form and function but not meaning that determine the classification of a word into a particular part of speech: 

The test of conversion is that the converted word adopts all the formal characteristics (inflection, etc.) of the part of speech it has been made into. Thus walk in he took a walk is a noun because it takes the form-word the (sic) before it, because it can take a plural ending -s, and so on. The question, which part of speech a word belongs to is thus one of form, not of meaning. The nouns in silk thread, gold watch are used as attribute-words very much as the adjective silken, but nevertheless they are not adjectives in the above collocations: we could not say *very silk, *more silk, as we could say very silken, more silken. (Sweet, 1891:39).  

Moreover, Sweet could be associated with the idea of the morpheme in the same way as Wrenn associated him with the idea of the phoneme. Even though he does not explicitly talk of it, there are many passages in his work where he makes use of the concept morpheme. Thus, for example, when he faces the problem of word-division, he affirms:  

It is evident that word-division implies comparison. As long as we confine ourselves to the examination of isolated sentences, we shall not advance one step further. But when we compare a variety of sentences in which the same sound-groups are repeated in different combinations, we are able first to distinguish between meaning and unmeaning sound-groups, and finally to eliminate a certain number of groups having an independent meaning and incapable of further division. (Sweet, 1876:11). (Our emphasis).  

What Sweet is presenting us here with is a kind of technique that could easily be compared to the structuralist techniques of discovery procedures. The idea of 'morpheme' as an abstract representation of grammatical categories was also advanced by Sweet: 

Sometimes an inflectional function is performed by a variety of distinct forms, as in the plurals trees, children, men and the preterites called, thought, saw, held. As the change of child into children has exactly the same meaning as that of tree into trees, we do not hesitate to regard all these changes as constituting one and the same inflection, however distinct they may be in origin, and so also with the preterites called, thought, etc. It sometimes even happens that different words stand in an inflectional relation to one another, with or without the help of inflection. Thus went, was stand in the same relation to goes, is as called, saw to calls, sees. (Sweet, 1891:29).  

Sweet's analysis of words is very similar to that of American structuralist grammarians. Sweet starts by studying what he calls words and structuralists stems, and then he proceeds with the analysis of derivational suffixes and prefixes, to finish with inflections. The two following quotations by Sweet are a good reflection of how close he was to the formulation of the morpheme:  

The inflected words cats can be divided into cat-s, but the second element, though it has the definite meaning of plurality, is not an independent sense-unit, and the connection between cats and the uninflected cat is so intimate that we cannot regard the two as distinct words. (Sweet, 1891:20).  

Such a derivative element as un- in un-known is an ultimate sense-unit with a very definite meaning, being so far on a level with the word not. But as it is not independent ... un- cannot stand alone, and can be used only with certain words. (Sweet, 1891:26-27).  

But Sweet's formal approach goes further than this. Another important tenet for American structuralists is the search of structural meaning as the most important task of any grammatical analysis. Fries affirms: 

The total linguistic meaning of an utterance consists of the lexical meanings of the separate words plus ... structural meaning ... Structural meanings are not just vague matters of the context, so called; they are fundamental and necessary meanings in every utterance and are signalled by specific and definite devices. It is the devices that signal structural meanings which constitute the grammar of a language. The grammar of a language consists of the devices that signal structural meanings. (Fries, 1957:56).  

For American structuralists, the general types of devices that English has to express structural meaning are the use of form-words (inflections and derivations), the use of function words (prepositions, determiners, subordinators, etc.), the use of word order, and in some cases the use of stress and intonation. Sweet's classification of parts of speech is also based on the same criteria:  

There are five ways of indicating the relations between words in word-groups and sentences: (a) word-order, or position, (b) stress, (c) intonation, (d) the use of form-words, and (e) inflection. (Sweet, 1891:30).

  As regards word-order and position, Sweet claims: 

The simplest and most abstract way of showing the relations between words is by their order. We see how the meaning of a sentence may depend on the order of its words by comparing the man helped the boy with the boy helped the man ... (Sweet, 1891:31).  

For Sweet, position is an essential criterion for his classification of parts of speech. An example of this is the following quotation: 

It is often difficult to draw the line between adjective-pronouns and ordinary adjectives. But if an adjective does not show any of the above formal peculiarities, it cannot be regarded as a pronoun, however much it may resemble an adjective-pronoun in meaning. Thus several is a pronoun because it can be used absolutely, as in I have several; but although divers has the same meaning as several, we cannot say *I have divers any more than we can say *I have good in the sense of I have good books; so divers can be regarded only as an ordinary adjective. (Sweet, 1891:70).  

As regards stress and intonation, Sweet affirms: 

English uses word-stress to express differences of meaning [as] in such pairs as 'abstract and abs'tract ... Stress and intonation, however, have not much influence on the grammatical structure of sentence. (Sweet: 1891:31).  

In relation to the role of inflections in Modern English, Fries states: 

In the common school grammars of English, inflections or the forms of words have received the major emphasis, and those matters of structure which did not parallel the devices of Latin have received very little or no treatment ... In English, inflections or the forms of words have tended to disappear as a grammatical device until in Present-day English the only really live uses of the forms of words to express grammatical ideas are (a) those to distinguish plural and singular number in substantives and (b) those to distinguish past and present tense in verbs ... (Fries, 1940:108-109).  

Sweet entirely agrees with Fries' words. In his analysis of English inflections, Sweet starts with case. After making explicit that English has only two case distinctions, he concludes: 

When we consider that the genitive inflection can generally be replaced by the preposition of, we see to what narrow limits the English cases, or rather case, are confined. The verbal inflexions are hardly less limited. The only personal inflexion is the s of he goes, which is practically a superfluous archaism. The only other inflexions are those which form the preterite and the two participles. These, together with the plural of nouns, are the only essential inflexions of English. (Sweet, 1876:19).  

As regards function words, Sweet also realizes that there are some words which express grammatical meaning: 

In English, verbs are modified partly by inflection, partly by form-words - particles [to] and verbs - which latter constitute the periphrastic forms of the verbs ... The form-words used to modify the English verb are called auxiliary verbs, or auxiliaries. (Sweet, 1891:88).  

In this respect, there is another point we would like to emphasize. As regards parts of speech, one of the greatest achievement of American structuralists was the sharp distinction between content and function words. This distinction is also clearly made by Sweet. Sweet treats it from three different point of views: the semantic, the formal and the functional. Semantically, Sweet distinguishes between full-words and half-words:  

There is an intermediate class of sound-groups, which, although not capable of being isolated and forming sentences by themselves, are yet not utterly devoid of meaning, and can, therefore, be to a certain extent isolated in thought, if not in form. Thus, if we compare the three groups (maen), (amaen) and (dhemaen), we see that the two prefixes have an unmistakable, though somewhat vague meaning of their own, which enables us to identify them at once in all other cases in which they are prefixed to nouns, and yet these two syllables would convey no meaning if pronounced alone ... In the case of (amaen) and (dhemaen) ... it seems best to distinguish two classes of words, full-words and half-words, (maen) being a full-word, (dhe) a half-word - that is, a word incapable of forming a sentence by itself, or of suggesting an independent meaning. (Sweet, 1876:11-12).  

Functionally, the distinction is based upon the following contrast: 

In such a sentence as the earth is round, we have no difficulty in recognising earth and round as ultimate independent sense-units expressing the two essential elements of every thought - subject and predicate. Such words as the and is, on the other hand, though independent in form, are not independent in meaning: the and is by themselves do not convey any ideas, as earth and round do. We call such words as the and is form-words ... [is] serves to connect subject and predicate ... then ..., though it has no independent meaning, has a definite grammatical function - it is a grammatical form-word. (Sweet, 1891:22). 

And formally, Sweet distinguishes: 

The parts of speech in inflectional languages are divided into two main groups, declinable, that is, capable of inflection, and indeclinable, that is, incapable of inflection. The declinable parts of speech fall under the three main divisions, nouns, adjectives and verbs ... Pronouns are a special class of nouns and adjectives, and are accordingly distinguished as noun-pronouns, such as I, they and adjective-pronouns, such as my and that in my book, that man. Numerals are (36) another special class of noun and adjectives: three in three of us is a noun-numeral, in three men an adjective-numeral ...  

Indeclinable words or particles comprise adverbs, prepositions, conjunctions, and interjections. The main function of adverbs, such as quickly and very, is to serve as adjunct-words to verbs and to other particles, as in the snow melted quickly, very quickly ... Prepositions are joined to nouns to make them into adjunct-words ... Conjunctions are used mainly to show the connection between sentences ... (Sweet, 1891:36-37).  

But Sweet does not stop here. For him, the distinction between indeclinable and declinable words is not simply a question of form: 

The distinction between the two classes which for convenience we distinguish as declinable and indeclinable parts of speech is not entirely dependent on the presence or absence of inflection, but really goes deeper, corresponding, to some extent, to the distinction between head-word and adjunct-word. The great majority of the particles are used only as adjunct-words, many of them being only from-words, while the noun-words, adjective-words and verbs generally stand to the particles in the relation of head-words. (Sweet, 1891:38).  

It should be clear by now that Sweet's reaction to traditional grammar is very similar in terms to that of American structuralism. However, there are two important points that cannot be ignored. In the first place, Sweet's reaction anticipated that of American structuralists by more than fifty years, and, secondly, Sweet's position has not been so widely recognized, and therefore it is not so popular.  

            A tentative corollary of all this could be the lack of information that American linguists had of the work which was produced in Europe.^[5] This lack of communication has been pointed out by many linguists. Gleason tries to explain the reasons for this isolation: 

More regrettable has been the isolation of American descriptive linguistics from parallel developments in Europe. Partly this is the result of differences in the academic roots. American linguists has been closely associated with anthropology ... European linguistics has had much less close connections with anthropology, and its moulders have been largely trained in Indo-European historical linguistics, classical and modern European languages, or literary criticism ... Another factor in the isolation has been the heavy concentration of American work on North American Indian languages. European workers have been more concerned with Old World languages, with the heaviest concentration on those of Europe and classical antiquity. Working on the same or closely related languages can be a strong force to bring linguists together. (Gleason, H.A., Jr., 1970:211-212).

  But whether from isolation or any other cause, we think that it can be concluded that, although American structural linguistics contributed much to the field of linguistics in general and to the description of English in particular, there were other linguists who preceded them that contributed in a similar direction towards a better understanding of the way in which language in general and English in particular work. One of these men was undoubtedly Henry Sweet. We would like to finish this paper with some words by J.R. Firth. Talking of the development of English grammar Firth said: 

In some respects there was no improvement in the basic theoretical approach until Sweet treated the whole subject at a much higher level. (Firth, 1957:159).

REFERENCES  

Bloomfield, L. (1933), Language. New York: Henry Holt & Co.

Bolton, W.F. & D. Crystal (edd) (1969), The English language: essays by linguists and men of letters. 2nd vol. Cambridge: Cambridge University Press.

Firth, J.R. (1957), "Atlantic linguistics", in J.R. Firth, Papers in linguistics 1934-51. London: Oxford University Press, 156-172.

Fries, C.C. (1940), American English grammar. New York: Appleton-Century-Crofts.

Fries, C.C. (1957), The structure of English. London: Longmans. (1st ed. 1952, Harcourt, Brace & Co).

Gleason, H.A., Jr. (1965), Linguistics and English grammar. New York: Holt, Rinehart & Winston.

Gleason, H.A., Jr., (1970), An introduction to descriptive linguistics. London: Holt, Rinehart & Winston. Revised ed.

Henderson, R. (1982): The indispensable foundation: a selection from the writings of H. Sweet. London: Oxford University Press.

Hill, A.A. (1958), Introduction to linguistic structures: from sound to sentence in English. New York: Harcourt, Brace & World.

Leitner, G. (1986), "English grammars - past, present and future", in G. Leitner, The English reference grammar. Tübingen: Max Niemeyer, 409-431.

Leitner, G. (1991), "Eduard Adolf Maetzner (1805-1902)", in G. Leitner, English traditional grammars. Amsterdam: John Benjamins, 233-55.

Michael, I. (1991), "More than enough English grammars", in G. Leitner, English traditional grammars. Amsterdam: John Benjamins, 11-26.

Nida, E.A. (1966). A synopsis of English syntax. The Hague: Mouton. 2nd ed. (1st ed. 1960).

Sledd, J. (1959), A short introduction to English grammar. Glenview, Illinois: Scott, Foresman & Co.

            Sweet, H. (1876), "Words, logic and grammar", Transactions of the Philological Society. (References here are to its reprint in Bolton & Crystal (edd) (1969), 8-29).

Sweet, H. (1891), A new English grammar. Part I: introduction, phonology, accidence. Oxford: Clarendon Press. (References here are to the 1968 impression).

Sweet, H. (1898), A new English grammar. Part II: syntax. Oxford: Clarendon Press. (References here are to the 1971 impression).

Wrenn, C.L. (1967), "Henry Sweet", in C.L. Wrenn, Word and symbol: studies in the English language. London: Longman, 150-169.

rk are no

1.        Michael affirms that "one of the most curious features of the grammars is their abundance. In every year of the century there were produced, on average, between eight and nine new grammars intended wholly or partly for school use." (Michael, 1991:12).

2.           In Britain we can find, for instance, Latham's Elementary English grammar. For use in schools (1843) or Bain's Higher English grammar (1863/1879). Sweet also acknowledges some debt to Bain's work.

3.           Gleason (1965:36) affirms that "During the nineteenth century the attention of most linguists was focused on standard written languages. Only rarely were spoken languages observed." He also makes the following contrast between traditional and structural grammar: "The traditional grammars had been largely based on written English, and whatever was said about spoken English had been fitted into the matrix set by literary language. For Trager and Smith only speech is really the language; writing is merely a reflection of speech, and often quite imperfect at that." (Gleason, 1965:83).

4.           There are very few corpora of spoken English, and those we have are normally very small if compared to those of the written language. Whereas the LOB and Brown corpora (both based on written material) are about one million words, the S.E.C. (based on spoken English) is only about fifty thousand words. The London/Lund Corpus (based on spoken English) is larger, since it has about five hundred thousand words. The Birmingham (Cobuild) Corpus has only about one million words from spoken records out of the twenty million words it contains. Therefore, it can be concluded that less than ten per cent of the material now available is from spoken sources.