Our phonemes of speech perception turn out to be identical to the concept of phonemes developed by the Leningrad Phonological School (LPS). (Please allow me not to rename it to St. Petersburg. Not at all out of special love for Comrade V.I. Lenin, but because it was formed under this very name). The founder of this school, Academician Lev Vladimirovich Shcherba, worked in the first half of the 20th century in St. Petersburg - Petrograd - Leningrad. He and his students focused on the task of teaching foreign languages, setting the correct pronunciation. Most foreign language textbooks in their phonetic part use the concepts and terminology developed by Shcherba. Shcherba's phonological theory itself was best presented in his textbook Phonetics of the French Language. In the future, these same concepts were supported by researchers involved in instrumental studies. sound speech and designing automatic speech recognition systems.
Moscow Phonological School
The concept of speech production phonemes turns out to coincide with the phonological system according to the theory of the Moscow Phonological School (MPS). A prominent representative of this school is Alexander Alexandrovich Reformatsky. The main works in which the views of this trend are formulated are devoted to the description of the native (Russian) language. Initially, each phonological school considered its constructions as the only true doctrine of the sound structure of the language. In the course of time, however, mainly in the depths of the Moscow school, the tendency to discuss problems in a comprehensive manner and to synthesize phonological theories prevailed. Ruben Ivanovich Avanesov, one of the IDF founders, made the first attempt at such a synthesis. He put forward the concept of “weak phonemes”, which, along with “strong” ones, are part of linguistic signs. If the phoneme of speech perception is a set of indistinguishable sounds determined by the position in speech, the phoneme of speech production is a program for choosing one or another sound depending on the position, then Avanesov’s weak phoneme is a set of differential features (those and only those) that must be specified for definition of sound in this position. From the point of view of the structure of the linguistic mechanism, Avanesov's phonemes really occupy an intermediate position between the phonemes of speech production and speech perception. They are associated with commands to the executive organs of speech, developed by programs for the implementation of signs in order to create one or another acoustic effect corresponding to the required phoneme of speech perception.
Prague Phonological School
Another phonological theory, intermediate between the theories of LPS and MPS, was developed by the so-called Prague Phonological School (PPS), which arose in Prague simultaneously with MPS and LPS by the works of Russian linguists who emigrated from the revolution. It was this school that became most famous in the West, and its most prominent representative, Nikolai Sergeevich Trubetskoy, is considered the founder and classic of world phonology. Similarly to Avanesov, Trubetskoy distinguishes two types of sound units in the composition of a word - phonemes and archiphonemes. Archphonemes appear in those cases when the conditions of the speech chain do not make it possible to recognize which particular phoneme of speech production was the basis for the appearance of a given sound. The concept of an archphoneme essentially coincides with the concept of Avanesov's weak phoneme. Another interpretation of the phenomenon of neutralization of phoneme differences in the speech chain was given by the Moscow phonologist Pyotr Savich Kuznetsov in the concept of hyperphoneme. A hyperphoneme is the set of all phonemes that can give given sound. From the point of view of the structure of the language mechanism, such a unit corresponds to the development of a system of hypotheses regarding the comparison of the chain of phonemes of speech perception perceived by hearing with one or another sign (word) represented in memory by a chain of phonemes of speech production.
American phonologyIn the same years, at the beginning of the 20th century, a school of descriptive phonology developed in the United States, which solved the problem of describing the languages of the American Indians. Their concept was close to the views of the Leningrad phonological school. In particular, the American dicryptivists most clearly formulated the procedure for dividing the speech stream into phonemes of speech perception. In the post-war years, under the influence of the advances in computer technology, American linguists for the first time directly raised the question of the technical modeling of language ability. The pioneer of these works was also a native of Russia (or rather from Poland) Naum Chomsky (Americans pronounce this name as Noum Chomsky). His work founded the direction called generative linguistics. Its task was set as the task of constructing a formal model (automaton) for the production (generation) of correct statements in a particular language. The phonological part of the generative theory arose thanks to the work of another Russian, Roman Osipovich Yakobson, who, in connection with the Second World War, emigrated from Prague (where he was a prominent member of the Prague School) to America. Describing the generation (production) of speech, generative phonology naturally came to a concept close to the Moscow phonological school. True, it must be said that at first the generativists tried to interpret the production of speech too abstractly as an action of some kind of formal calculus, like algebra, which, however, led to the emergence of the theory of formal languages within the framework of mathematics, which already has an indirect relation to linguistics. The general scheme of phonetic speech production in generative phonology is that linguistic signs, through successive transformations according to language rules, are transformed from an internal (deep) representation in the phonemes of speech production into a surface representation by speech sound types. Accepting the terminology of generativists, we can call the phonemes of speech production - deep phonemes, and the phonemes of speech perception - surface phonemes.
The essence of the phonological theory underlying the present description of the phonological system of the Russian language can be briefly expressed in six formal conditions that any phonological description must satisfy.
As I progress, I will dwell on the meaning of these conditions, and therefore on the proposed theory for describing purely phonetic facts. I will try to compare the implications of this theory with those of other phonological theories.
The completely exhaustive and generalized nature, as well as the simplicity of the "practical" consequences arising from the proposed theory, determine its expediency.
1.1. Condition (I). In phonology, speech facts are presented as sequences of units of two types: segments, to which certain phonetic features (both articulatory and acoustic) are attributed, and boundaries, which are characterized only by the influence they exert on the segments.
1.2. Condition (2). The phonetic features used to characterize the segments belong to a special, narrowly defined class of features called distinctive. All distinguishing features are binary.
Acceptance of condition (2) entails the description of all segments in all languages in terms of a limited list of properties, such as "nasality", "sonority", "softness", etc. With respect to this list of properties, only question: "Does this segment have a certain property?" It follows that the differences between segments can only be expressed by the differences between which distinguishing features are included in one segment and which in another. Therefore, segments (even in different languages) can differ from each other only by a limited number of differences.
Most linguists and phoneticians believe that all human languages can be characterized by a limited number of phonetic features. This point of view is expressed in one form or another in many works on general phonetics, from Bell's Visible speech, published in 1867, to Heffner's General phonetics, published in 1949. However, many scholars do not agree with this opinion. They believe, in the words of one of the authors, that "languages can differ from each other indefinitely and in the most unexpected way."
It follows from this that condition (2) and the point of view given above are mutually contradictory judgments about the nature of human language and are subject to empirical verification. If the study of the most diverse languages showed that the number of different phonetic features necessary for a phonological description increases with the number of languages studied, then condition (2) would have to be rejected. If, on the contrary, such an analysis would show that as more and more languages are included in the study, the number of different phonetic features slightly exceeds or does not exceed at all some finite small value, then condition (2) should be accepted.
Despite the fact that languages have been discovered that have phonetic features that are not inherent in Western
Languages, the number of such features should not be exaggerated.
By studying phonetic models that have been tested in many languages, for example, the models described in the book by N. Trubetskoy "Grundzuge der Phonologie" or in the book by C. L. Pike "Phonetics", as well as the modified international phonetic alphabet 1PA, successfully used in England in the study of African and Oriental languages, one cannot but pay attention to the small number of phonetic features encountered (on the order of twenty or less). Since the languages described represent a very significant part of all the languages of the world, it can be expected that the number of relevant phonetic features will not increase significantly as all new languages become subjected to scientific research. Therefore, it seems that there are no sufficient grounds to reject condition (2) for this reason.
On the other hand, condition (2) entails an even stricter restriction. It requires segments to be defined using a small number of binary properties: distinguishing features. Systematic examination of the available material in various languages has demonstrated the complete suitability of the binary distinguishing feature model for phonological description. Until now, no examples have been given that would cast doubt on the correctness of the binary scheme. On the contrary, the extension of the binary structure to all glyphs made it possible to obtain a satisfactory explanation of some "incomprehensible" phonetic changes and made it possible to formulate a methodology for evaluating phonological descriptions.
1.8. Segments and boundaries are theoretical constructs. Consequently, they must be appropriately correlated with observable objects, that is, with the actual facts of speech. The weakest condition imposed on phonological description and accepted by all is
Condition (3). The phonological description must provide a method for obtaining (extracting) the original utterance from any phonological notation without referring to information not contained in this notation.
In other words, it is assumed that it will be possible to read a phonological notation regardless of whether its meaning, grammatical structure, etc. are known. Obviously, this will be achieved only when all distinct statements are written in different sequences of characters. However, it is not at all necessary to fulfill the opposite requirement, since it is possible to write rules that provide for the same reading of several non-identical sequences of characters. For example, the character sequences (m'ok bi) and (m'og bi) would be pronounced the same if a rule were formulated according to which unvoiced consonants would be voiced before voiced consonants. However, in this case it will be impossible to determine from the statement alone which of the two (or more) sequences of symbols is the actual representation of the given statement. Thus, in the example above, the listener of an utterance would not be able to select either of the two phonological representations of that utterance unless he consulted the meaning or other information not contained in the signal. It follows that a given sequence of sounds must be represented by only a single sequence of symbols. Only in this case will phonological descriptions satisfy:
Condition (For). The phonological description should include the rules for obtaining (extracting) an accurate phonological representation of any speech fact without resorting to information not contained in the physical signal.
1.31. There is the simplest way to construct a phonological description that would satisfy condition (3a). This method consists in creating such a system of symbols, in which each symbol will correspond to one sound and vice versa. If the symbol system is exhaustive in the sense that it contains a symbol for any sound, then every person familiar with the phonetic significance of symbols will be able not only to correctly read any sequence of symbols, but also to write down any utterance unambiguously in the form of a corresponding sequence of symbols. It was in this way that the phoneticians of the end of the last century tried to construct a notation that satisfies the condition (3a). This is reflected in the famous slogan of the "International Phonetic Association" "Association Internationale de Phonetique"): "For everyone
/sound is a special character. However, it is well known that all attempts to put this idea into practice were unsuccessful, since they inevitably led to a seemingly infinite increase in the number of symbols, for, strictly speaking, two identical sounds do not exist. The only reasonable way out of this situation would be some limitation on the number of characters.
1.32. This idea can be formulated as:
Condition (3a-/): Different statements only
must be written using different sequences of characters. The number of distinct characters that are used in all records required for this purpose should be kept to a minimum.
In other words, the requirement "for each sound - a special symbol" was replaced by the requirement "for each utterance - a special notation", and a limit was imposed on the number of symbols used for the notation. However, the restriction caused a number of difficulties. For example, in English [h] and do not occur in the same environment. According to condition (3a-1), they should be considered as positional variants of one phoneme, which strongly contradicts our intuitive idea. Even more amazing is the fact that any number (of facts, statements, people) is always possible to represent as a binary number. It follows from this that condition (3a-1) can be satisfied in a very trivial way, which is to assume an alphabet consisting of only two characters. This, however, can be done without considering the phonetic facts. Thus, one can come to the absurd conclusion that the number of phonemes in all languages is the same and equal to two.
In order to overcome these difficulties, it was proposed to consider the positional variants of the same phoneme as "phonetically the same". Unfortunately, this approach only postpones the resolution of the problem until the next stage, which is to answer the question of what is meant by the term "phonetically the same." Apparently, this is only a modified form of another question, to which there is still no answer: what is meant when they say that two sounds are the same.
1.33. Let us now consider what is the effect that condition (3a) has on the phonological recording of certain speech facts. In Russian, voiced
is a distinguishing feature of all noisy ones, except for /s/, /s/ and /x/, which do not have voiced correspondences. These three noisy consonants are always voiceless, except when they are followed by a voiced noisy. In this position, these consonants are voiced. However, at the end of a word (this is inherent in all Russian noisy ones), they become deaf, if the next word does not begin with a voiced noisy one, in which case they become voiced. For example, “would it be?”, Hofm’og bi] “would it be”; “whether to burn”, but “would burn”.
If we write down the above statements in phonological notation that would satisfy both condition (3) and condition (3 a), then they would look like this: /m'ok 1,і/, /m'og bi/, /z 'ec 1,і/, /z'ec bi/ . In addition, a rule would be needed stating that noisy ones that do not have voiced correspondences, i.e. /s/, /s/ and /x/, are voiced in a position before voiced noisy ones. However, since this rule is valid for all noisy ones, the only result of an attempt to fulfill conditions (3) and (3 a) will be to divide the noisy into two classes and establish a special rule. If condition (3a) is omitted, then the four statements can be written as follows: (m'ok l,i), (m'ok bi), (2'ec l,i), (z'ec bi), and the above the rule will be extended to all noisy ones instead of (c), (c) and (x). Thus, it is obvious that condition (3a) leads to a significant complication of the notation.
Traditional linguistic descriptions included both notation systems satisfying condition (3) only, and systems satisfying conditions (3) and (3a). The former were usually called "morphophonemic", in contrast to the latter, which were called "phonemic". In a linguistic description, one cannot do without morphophonemic notation, since only with its help it is possible to resolve the ambiguity arising from homonymy. For example, the fact that the English phonological notation (tacks "buttons" and tax "tax") is ambiguous is usually explained by the morphophonemic difference between "phonemically identical" utterances.
Note, however, that for the examples from the Russian language discussed above, the morphophonemic notation and the rule regarding the distribution of voicedness are quite sufficient for a satisfactory description of the actual facts of speech. Consequently, phonemic recording systems constitute a certain additional level of displaying the facts of speech, the need for which is due only to the desire to fulfill the condition (3 a). If condition (3a) can be omitted, then the need for a "phonemic" notation will also disappear.
1.34. Condition (3 a) refers to essentially analytic operations. Analytical operations of this kind are well known in all sciences. Quantitative and qualitative analysis of chemical compounds, analysis of electrical circuits, identification of plants and animals, medical diagnosis can serve as examples of finding appropriate theoretical models for various complexes of observed objects (for example, respectively, the chemical formula, the location of the circuit elements, the classification within the general scheme, the name of the disease ). However, the theoretical constructs that make up the models that arise as a result of various analytical operations are postulated within individual sciences without taking into account the operations by which these constructs are abstracted from observable objects. The introduction of theoretical constructs is never based on considerations related to analytic operations.
So, for example, it is unimaginable that in chemistry substances that can be determined visually stand out in a separate class, in contrast to substances that require the use of more complex methods for your definition. However, this is precisely the meaning of condition (3a), since it introduces a distinction between phonemes and morphophonemes, proceeding only from the fact that phonemes can be determined on the basis of only acoustic information, while morphophonemes require additional information for their definition.
Such a serious deviation from the usual scientific practice can only be justified if it can be shown that the differences between phonology and other sciences are so great as to make such a deviation necessary. However, this has not been proven so far. On the contrary, it is quite common to try to emphasize the internal identity of the problems of phonology and other sciences. This leads to the conclusion that condition (3a) is an unjustified complication that has no place in scientific description language.
The elimination of condition (3a) does not run counter to the linguistic tradition to such an extent, as it might seem at first glance. It is hardly accidental that in the phonological descriptions of E. Sapir and, to some extent, JI. Bloomfield condition (For) is absent.
1.4. Condition (4): The phonological description must be appropriately included in the grammar of the language. Particular attention should be paid to the phonological notation of individual morphemes. This entry must be chosen so as to obtain simple rules for all grammatical operations in which morphemes can participate (for example, inflection and word production).
In this paper, grammar is considered as a way of identifying all the sentences of a language. Therefore, it can be considered a common definition of the term "sentence in the language L". In its structure, grammar resembles a system of postulates, from which, by applying certain rules, theorems can be deduced. Every sentence in a language can be considered a theorem of the system of postulates that makes up the grammar.
The identification process begins with the symbol "Sentence", since it is this term that is subject to explication through grammar.
In the process of identification, the indicated symbol is transposed into various notation systems related to each other by means of certain rules; at each stage of identification, the sentence is written with a certain combination of symbols (not necessarily their one-dimensional sequence), which is a consequence of the application of grammar rules. In order to delimit individual characters from each other and connect them with neighboring characters, each character is preceded and followed by a special sign &. It will be shown below that these signs play an important role in the phonological notation of a sentence, since some of them are eventually transposed into phonological boundaries. The last stage of sentence identification is the transposition of the abstract notation into sound.
The transposition rules that make up the grammar can be generally represented by the formula: "replace x with y under the condition z". However, the rules differ from each other in the type of record that results from applying each of them.
The differences in record types are the result of restrictions placed on the possible values accepted by variables x, y and z. The set of rules that generate a particular type of record is called the linguistic level.
The purpose of applying the rules of the highest level, the so-called level of immediate constituents, is to obtain tree-like models that represent the structure in terms of the immediate constituents of the sentence. An example of such a tree is the partial structure of a Russian sentence presented on p. 309 in terms of direct constituents.
The structure of a sentence in terms of direct constituents is considered fully identified when, applying the rules listed, none of the symbols can be replaced by another (for example, the symbol & The substantive group & cannot be replaced by any other symbol from the rules listed above).
These "non-replaceable" characters are called terminal characters, and the sequence of such characters is called a terminal string. However, since there are many more rules in the current grammar than in our example, the "irreplaceable" characters in it are not in fact the terminal characters of the grammar rules of the Russian language itself.
The rules that generate this tree are as follows:
Replace & Offer & with & Adverb & Subject- & Say-
grouping
„ & Adverb & to & Adverb & -(2)
„ & Subject & on & Substan- & Im. n. & - (3)
„ & Predicate & on & Verb- & -(4)
new group
„ & Verb- & on & Verb transition. & Extra- & -(5)
new group
„ & Verb trans. & on & Prefix & Stem & Past.
verb tense & trans. - (6)
„ & Supplement & on & Substance- & Vin. n. & - (7)
Different branching points of the tree correspond to different direct constituents of the sentence. Therefore, the tree displays the structure of the sentence
by its immediate constituents, and the grammar rules of the immediate constituents are a formal analogue of the analysis by its immediate constituents. In order for the rules to generate trees of this type, it is necessary to restrict them so that one rule cannot replace more than one character. This restriction also provides for the obligatory finding of the tree of the syntactic structure for each terminal string. In addition, it becomes possible to choose a completely unambiguous path from the initial symbol & Sentence & to any other symbol (immediately constituting) of the tree. This path is called the derivational history of the symbol.
Next, the transformation level rules are applied to the trees. At the transformational level, one rule can replace more than one character. This allows changes to be made to the notation that would not have been possible with the application of the grammar rules of the constituents themselves. For example, you can change the order of characters in a sequence or exclude some characters altogether. In addition, the transformation rules take into account the derivational history of individual characters. Therefore, it becomes possible, for example, to formulate different rules for generating the symbol & Substantive group & from the symbol & Subject &, on the one hand, and for generating the same symbol & Substantive group & from the symbol & Object &, on the other hand. It is believed (due to the involvement of the derivational history of individual symbols) that the transformation rules apply to the structure trees of the immediate constituents, and not to the terminal strings.
The last set of rules, the so-called phonological rules, provides for operations on transformed terminal strings, consisting exclusively of special kinds of segments and boundaries. Operations consist in the final attribution of phonetic features to the segments.
In contrast to the grammar rules directly constituting one phonological rule can replace more than one character. However, the phonological rules do not take into account the derivational history of the symbols on which operations are performed.
1. 41. Up until now, we have written sentences only with symbols representing a certain class of morphemes, for example: & Subject &, & Adverb &, & Im. etc. Obviously, at some stage in the process of defining a sentence, these morpheme class symbols must be replaced by actual morphemes; for example, the symbol & The adverb & must be replaced by one or another adverb of the Russian language. This substitution can be made at the level i of direct constituents, which is carried out by applying the rules of the type:
„replace & Adverb &
where A, B, C denote the corresponding Russian adverbs, for example, such as there, quickly, yesterday, etc. Rules of this type make up the dictionary of the language.
The choice of some morphemes is determined by the context in which they occur. For example, in Russian there is a close relationship between the phonological composition of the morpheme that replaces the symbol & The stem of the verb &, and the choice of the present tense suffix.
In principle, it is possible to argue about what is definable in this case, and what is defining. However, in all the cases that I have had occasion to study, considerations of elementary economy require that the choice of the suffix depend on the choice of the stem, and not vice versa.
Such considerations have always underpinned linguistic descriptions and have been useful in distinguishing between lexical and grammatical morphemes. In this paper, it is not possible to delve into the question of which classes of morphemes are lexical and which are grammatical. For our purposes, it suffices to establish that such a distinction is necessary and that lexical morphemes must be introduced into phonological notation before grammatical morphemes.
1.42. Let us now consider how individual grammatical morphemes are introduced into phonological notation. The application of the grammar rules of the immediate constituents, which up to now have satisfactorily served our purposes, leads in a number of cases to difficulties. Let's consider these cases. In Russian, it happens that & Noun & is a homophone of & Adjective &; for example, (s, n,) "blue" as & Noun & and (s,'in,) "blue" as & Adjective & in the meaning of "blue". Moreover, both & Noun & and & Adjective & are used before grammatical morphemes of the same class, for example, before & Mn. h. & Im. P. &. Therefore, according to the rules of grammar of the immediate constituents and & Adjective &Mn. & Them. n. & and & Noun & Mn. h. &, Im. n. & must give & (s,’in,) & Mn. h. & Im. P. &. Here a significant difficulty arises: &Mn. h. & Im. n. & is carried by different suffixes depending on what it follows, i.e., whether it comes after & Noun & or after & Adjective &. However, according to the rules of the grammar of direct constituents, operations depending on the derivational history of symbols cannot be applied to a given sequence of characters. Therefore, it is impossible to transpose the sequence & (s,4n,) & Mn. h. & Im. n. & in two entries, i.e. in (s’,in, -і) “blue” in the case of & Noun & and (s,*in,-iji) “blue” in case of & Adjective &.
One can apparently get out of this difficulty by establishing additional grammar rules for the immediate constituents of the type:
Replace & Adjective & Mn. h. & Im. n. & on & Adjective & Mn. h. & Im. n. with lag. &
Replace & Noun & Mn. h. & Im. n. & on & Noun & Mn. Ch. & Im. n. creatures. &.
The above rules eliminate the ambiguity contained in the restrictions that operate at the level of direct components. However, one has to pay too dearly for this: the number of classes of grammatical morphemes increases. Instead of dealing with one class of grammatical morphemes & Im. etc. &, you have to break it into smaller classes, and the number of these classes will be very large, since homophony is observed not only between & Noun & and & Adjective &, but also between other classes.
Along with the difficulties that arise at the level of direct constituents due to the fact that several suffixes correspond to one class of grammatical morphemes, one has to meet with difficulties of a different kind that arise in connection with the very common phenomenon of "syncretism". In linguistics, the term "syncretism" refers to the phenomenon in which one symbol expresses several grammatical categories, for example case endings nouns in Russian, along with the case, usually indicate the number or gender. However, the rules at the level of immediate constituents are very strict in requiring that a single rule replace no more than one character.
Therefore, at this level it is impossible to apply a rule like: “replace & Mn. h. & Im. n. & to & (i) & ”, where two characters are immediately replaced - & Мнч. & and & Im. P. &. As a result, we can say that the morphological process of inflection cannot be included in the grammar rules of the direct constituents.
A natural solution to these difficulties is the inclusion of morphology (that is, that part of the grammar that considers the replacement of symbols of entire classes of grammatical morphemes by individual grammatical morphemes) at the transformational level, at which the two restrictions mentioned above lose their force. Such a solution seems to be especially expedient, because it coincides with the traditional method of studying the processes of morphology, in which over various
individual morphemes produce different
operations depending on which class these morphemes belong to. In traditional descriptions, replacing multiple characters with a single rule is common.
1.5. As already noted in § 1.41, the level of immediate constituents should contain rules of the type:
Replace & Adverb & with there (8a)
Replace & Adverb & with yesterday (86)
Replace & Adverb & with so (8c), etc.,
i.e. lists of morphemes. However, in the scientific description of a language, one cannot be content with compiling lists of all existing morphemes. Just as the syntax of a language is much more complex than an exhaustive list of all sentences, so the phonological description of a language is not a simple list of morphemes. A phonological description must include a statement of structural principles, of which real morphemes are special cases.
The process of generating a given sentence involves the selection of specific morphemes that make up the sentence from a number of possible options, i.e. from lists similar to rules (8a) - (8c). The choice of certain morphemes is carried out on the basis of extra-grammatical criteria. The grammar should provide rules for choosing one morpheme from the list, and these rules are introduced into the grammar from the outside (perhaps by the speaker himself). The rules must be given in the form "choose rule (8a)", and the grammar interprets them as a command to replace the symbol & Adverb & with there.
Instead of writing rules in an arbitrary numerical code that does not contain any information about the phonetic structure of morphemes, one can use for this purpose the notation of morphemes directly in terms of distinguishing features, which is much more consistent with the goals of linguistic description. So, for example, instead of the command “choose rule (8a)”, the grammar can be given the following command: “replace & Adverb & with a sequence of segments in which the first segment contains the following distinguishing features: non-vocal, consonant, non-compact, high tonality, non-tension, non-nasality etc.; the second segment contains distinguishing features: vocality, non-consonance, non-diffuseness, compactness, etc., and the third segment contains distinguishing features: non-vocal, consonant, non-compact, low tonality, non-tension, nasality, etc.”
It is convenient to represent such commands in the form of matrices, in which each vertical column contains one segment, and in each of the horizontal rows there is one distinguishing feature. Since the features are binary, the (+) sign means that the given segment has the given distinguishing feature, and the (-) sign means that the given distinguishing feature is absent. Such a record is shown in Table. 1-1 (see page 321).
Since the goal of the teams is to choose one morpheme from the list, the distinguishing features and their complexes, which serve to distinguish morphemes, will play an important role in the teams. Distinguishing features of this type and their complexes are called phonemic. Features and complexes of features that are distributed in accordance with the general rule of the language and, therefore, cannot serve to distinguish morphemes from each other, are called non-phonemic.
Each phonemic feature in the segment denotes some information introduced from outside. If grammar, as understood in this paper, reflects the actual functioning of a language, then the commands for choosing individual morphemes can be considered to be carried out by a conscious effort on the part of the speaker, in contrast to the fulfillment of various mandatory rules of the language, which the speaker of this language obeys automatically. Since we speak quite quickly, sometimes with an identification speed of up to 30 segments per second, it is reasonable to assume that all languages are constructed so that the number of distinguishing features identified when choosing individual morphemes does not exceed some minimum value. This assumption is expressed in the following formal requirement:
Condition (5). The number of identifiable distinguishing features used in a phonological notation should not exceed a certain minimum value necessary to fulfill conditions (3) and (4).
In the course of the further presentation, we will also operate with non-phonemic features that remain unidentifiable in the phonological notation. Such unidentifiable features will be conditionally denoted by zeros in the corresponding place of the matrix . Zeros are auxiliary characters used only for convenience of presentation; they have no function in the phonological system of the language.
1.51. Some features are non-phonemic because they can be predicted from some other feature in the same segment. So, for example, in the Russian language, the feature "diffusion - non-diffuse" is non-phonemic in relation to all sounds except vowels, i.e., it is possible to predict the distribution of this feature in all segments that are non-vocal and (or) consonant. Similarly, in segment (c), palatalization can be predicted in all cases, regardless of context.
In addition to cases of non-phonemic features that do not depend on the context, in all languages there are cases of non-phonemic features of individual features from separate segments that are included in special contexts. Since the application of condition (5) is not limited to individual segments, a feature must remain unidentified in a phonological notation if it is non-phonemic due to being used in a particular context. Such contextual constraints are called distributional constraints. Consequently, with the help of condition (5), distributive restrictions are introduced as an integral part of the grammar of the language. This is a great achievement of the present descriptive scheme, since the study and description of distributive restrictions presented significant difficulties in linguistic theory.
The following examples illustrate the place of distributive constraints in the present theory.
Example 1. Although combinations of two vowels are very common at the junction of morphemes, only two combinations of vowels are allowed inside a morpheme in Russian (*/ *i\ or (*a*i), for example (pa'uk) "spider", jkl / auz +a) "slander", (t,i'iinj "tiun". Thus, if it is known that the sequence of segments within the morpheme consists of two vowels, we will know in advance all the distinguishing features of the second vowel, except for stress, and all the distinguishing signs of the first vowel, except diffuseness and stress.Therefore, in the dictionary entry of a lexical morpheme containing such a sequence, it is necessary to indicate only the signs of vocality - non-vocality, consonance - non-consonance, stress - non-consonance, and for the first vowel also diffuseness - non-diffuseness. All other signs can be are uniquely predicted; therefore, according to condition (5), they must remain unidentified
Example 2. Within a morpheme, the voiced feature is not distinctive in front of noisy consonants, with the exception of (*y), followed by a vowel or sonorant, i.e., a nasal consonant, smooth, or glide. The voicedness or unvoicedness of a noisy sequence is uniquely determined by the last noisy of this sequence. If this noisy one is voiced, then the rest of the noisy ones are also voiced, if it is deaf, then the rest are noisy, respectively, deaf. This means that in such sequences the sign of sonority is undetectable for all noisy ones, except for the last one.
(*p *s *k) voiced 00 -
1.512. Cases where a feature can be predicted from grammatical context rather than from purely phonological factors are not, strictly speaking, distributive constraints. For example, in Russian there are nouns, some forms of which are characterized by the presence of stressed vowels, and others by the presence of unstressed vowels. For example, for a noun (v*al \ "shaft" in all singular forms, the stress falls on the vowel of the root, and in all plural forms - on case endings.
Thus, when writing the lexical morpheme (v*al\ in the dictionary, it is completely wrong to indicate that the root vowel is stressed. It would also be incorrect to indicate that the root vowel is unstressed. Strictly speaking, the sign of stress cannot be determined until the grammatical context in which (v*al\) is used is not known. However, as soon as this context becomes unnatural, the stress will be assigned automatically, according to the rules of declension of nouns. Since in this case the stress sign can be predicted based on other characters, which or otherwise must be present in the record, condition (5) requires that this feature be unidentifiable.
In cases where a feature is derived only from certain grammatical contexts, it is necessary to resort to a different description procedure. So, for example, in Russian, the appearance of a sign of voicedness in noisy consonants at the end of a word depends on whether the noisy one is voiced (except-(*£>)) or deaf. According to this rule, one can predict the voiced feature in the last segment of the word (r*og) "horn" in the nominative singular and in the accusative singular, but not in the other cases. Therefore, when writing this lexical morpheme, it is necessary to indicate the sign of the sonority of the last noisy consonant.
1.52. In Russian, there are a number of stems, the forms of which can have a fluent vowel. Wherever these "alternations cannot be predicted from other (i.e., grammatical or phonological) factors, they should be indicated when the morpheme is entered into the dictionary. This is done by inserting a symbol at the place in the word where the fluent appears. vowel, for example: jt'ur#k) "Turk", but (p'arkj- "park"; cf. the corresponding forms of them. singulars (t'urok) and (p'ark) and gender. (t'urk+a) and (p'ark+a) units.
Klagstad showed that, with a few exceptions that can be listed separately, vowel signs # can be determined from the context. Therefore, # can be characterized by signs of vocality and non-consonance; instead of other signs there will be zeros, i.e. # is a vowel without indicating the distinguishing features of vowels.
So, lexical morphemes are recorded in the dictionary in the form of two-dimensional tables (matrices), in which the vertical rows correspond to segments, and the horizontal rows correspond to distinctive features. Since all signs are binary, they are identified by plus or minus. Wherever a feature can be predicted from the context, this is reflected in the entry - the corresponding places in the matrix remain unidentifiable. Table I-1 shows a similar record of the sentence, the analysis of which at the level of direct constituents was given in § 1.4.
1.53. Now it is necessary to study in more detail the types of segments that can be included in matrices representing various morphemes. Let us define the following order relation between segment types: we will assume that the type of segment (L) differs from the type of segment (B) if and only if at least one feature that is phonemic in both types has a value in (L) different from from (5), i.e. plus in (L) and minus in (B), or vice versa.
Sign 1 + - + (L) "does not differ
Sign 20 + -
Feature 1 + - - All three types of segments
tov "various":
Feature 20+
The set of all types of segments that occur in matrices and represent morphemes of a language is called the set of fully identifiable morphonemes. Because fully identifiable morphonemes serve to distinguish one morpheme from another, they are analogous to "phonemes" and "morphophonemes" in other linguistic theories. We will write fully identifiable morphonemes in direct letters in curly braces (()).
Like other types of segments that occur in phonological notation, fully identifiable morphonemes obey condition (5), requiring that the number of identifiable features be minimal. It can be shown that imposing such a restriction on the set of fully identifiable morphemes is equivalent to requiring that the matrix consisting of the set of fully identifiable morphemes be represented as a tree. And if each branching point corresponds to a certain feature, and the two branches extending from each point represent the plus and minus values taken by the feature, then the path from the starting point to the terminal point of the tree will uniquely determine the fully identifiable morphoneme. Since such a diagram takes into account only phonemic,
T abl. I - 1. Recording the sentence given in § 1.4, following how the choice of lexical morphemes is given *
i.e., identifiable features, then fully identifiable morphonemes are uniquely determined by pluses and minuses, without taking into account unidentifiable features.
The possibility of displaying the matrix of distinguishing features in the form of a tree indicates the presence in the matrix of at least one feature identified in all segments. This feature corresponds to the first branching point and divides all types of segments into two classes. Each of the next two branch points corresponds to a feature identified in all segments of one of the two subclasses. These features may or may not be the same. Thus, all types of segments are already divided into four subclasses, with each of which the above operation can be done again, etc. If the subclass contains only one type of segments, this type is fully identifiable, and the path along the tree describes the composition of the distinguishing features of this type segments.
Thus, the representation of a matrix in the form of a tree is tantamount to establishing a certain hierarchy of features. However, such a hierarchy may not be complete. For example, if in the phonological system (see Table I-3) two features are fully identifiable, then any order in which these features are arranged will be satisfactory. A number of examples are analyzed below, of which the penultimate one illustrates the partial ordering of features according to various criteria. The existence of a hierarchy of features confirms our intuition that not all features carry the same weight in a given phonological system, for example, the distinction between vowels and consonants is more fundamental for various phonological systems than the distinction between nasal and non-nasal vowels or voiced and voiceless consonants.
The following examples depict matrices in the form of tree diagrams. Under some conditions, matrices can be represented as a tree, under others it is impossible. These conditions are discussed below.
Matrices of some types of segments cannot be represented as a tree. For example, the matrix shown below cannot be converted to a tree because it does not contain a fully identifiable feature (i.e., a feature that does not take the value "null").
On the left side of the tree resulting from this matrix, feature 2 precedes feature 3, and on the right side, feature 3 precedes feature 2.
I have not been able to establish whether such cases occur in natural languages.
Due to the fact that the ordering of features is free, it is possible to obtain several trees from one matrix that meet the above requirements.
In this case, when choosing one of these trees, one can be guided by condition (5), which gives preference to a tree with a more symmetrical shape. To illustrate, let's give an example from a particular system (similar to the phonological system of the Russian language), where various models are possible:
Obviously, the second model is more economical, since it contains more zeros, which is reflected in the greater symmetry of the second tree.
On fig. The I-1 phonological system of the language is represented as a "tree". The different paths along the tree from the first branch point to the terminal points define different fully identifiable morphonemes.
It will be shown below that segment types defined by paths starting from the first branch point and ending at intermediate points, i.e. segment types "not different" from several fully identifiable morphonemes, play an important role in the functioning of the language. We will call such types of segments incompletely identifiable morphonemes and designate them with asterisks for the corresponding fully identifiable morphonemes. It should be noted that a feature identified in a fully identifiable morphoneme may not be identified in an incompletely identifiable morphoneme only if all features located in the hierarchy of the tree below the given one are also unidentifiable.
1.54. It follows from condition (5) that only phonemic features are identified in the phonological notation. However, in a real statement there can be no unidentifiable features.
Languages differ from one another in the position they occupy in them. For some non-phonemic features, there are certain rules for their phonetic implementation, for others there are no such rules, and their implementation in each specific case depends on the speaker. It is this difference that underlies the opposition of the so-called allophones and free variants of phonemes.
Non-phonemic features as free variants cannot be properly included in a linguistic description. From the point of view of such a description, it is only of interest that they are free options. However, this information can also be transmitted simply by omitting any mention of the signs of interest to us. Thus, if the further description does not contain any information about the implementation of some feature in a particular context, this will mean that this feature is a free option.
1.55. The rules of grammar constitute some partially ordered system. Therefore, it seems quite appropriate to investigate what place in this hierarchy belongs to the rules that determine the non-phonemic distribution of features. In this paper, such rules will be called "F-rules". Recall that at the level of directly composing symbols of lexical morphemes, they are replaced by sequences of segments consisting of distinctive features (matrices). However, at this level, the symbols of the classes of grammatical morphemes remain unchanged in the notation (see Table I-1). Only after applying the transformational rules of inflection and word production, the symbols of the classes of grammatical morphemes (for example, "Past tense", "Singular", etc.) will be replaced by their phonological sequences, which are derived from them. Since transformation rules introduce additional feature segments into the record and also modify previously entered segments, placing F-rules before transformations may cause certain rules to be applied twice: once before the last transformation rules and a second time after the last transformation rule. So, for example, according to the transformational rules of declension of nouns in Russian, & (іb "ap) & Unit & Data. P. & is replaced by (iv" anu). If the rules for assigning non-phonemic features to unstressed vowels are applied prior to this transformation, then those same rules would need to be applied again during the transformation, or identify all non-phonemic features in () in some other way. Therefore, it seems most appropriate to place all the rules governing the distribution of non-phonemic features after the transformation rules. However, for a number of reasons, it is desirable that some F-rules be applied before transformations, even if this entails the difficulties described above.
For the Russian language, as well as for many other languages, the position is true, perhaps not universal, according to which, for the correct functioning of certain transformational rules, especially the rules of inflection and word production, it is necessary that certain features be identified in the record, regardless on whether these signs are phonemic.
So, for example, for the correct application of the rules of Russian conjugation, information is needed on whether the verb stem ends in a vowel sound. In the third segment of the stem of the verb “rva-t”, the signs “vocality - non-vocality” and “consonance - non-consonance” are non-phonemic, since in Russian in morphemes that begin with a sequence of segments, of which the first is smooth, and the second is a consonant, the third the segment must be a vowel (see § 2.161, rules of morphological structure; rule 1c). Thus, according to condition (5), the phonological record of the considered morpheme should look like this:
vocality - non-vocality + - O
consonance - non-consonance + + O
However, since the distinguishing features of the third segment remain unidentifiable, it cannot be established whether this segment is a vowel. Therefore, it is impossible to determine the correct conjugation of this verbal stem. However, if the F-rule, according to which these non-phonemic features are identified (morphological structure rule 1c), is applied before transformation, then these difficulties are easily eliminated. Since this example is no exception, we concluded that at least some of the F-rules should be applied before the transformation rules, regardless of the difficulties involved.
1.56. The considerations discussed above led us to the conclusion that it is necessary to divide all F-rules into two groups. One group includes morphological structure rules (MS-rules) that must be applied before transformations, the other group includes phonological rules (P-rules) that are applied after transformations. Naturally, the question arises of how to determine which F-rules are included in the MS-npa rules group and which are in the P-rules group. For the Russian language, the following criterion is quite satisfactory.
The rules of morphological structure must ensure that all segments appearing in a record are either fully or incompletely identifiable morphonemes.
In other words, the set of segment types resulting from applying the rules of morphological structure is determined by all possible paths along the tree, starting from the first branching point. As noted in § 1.53, this limits the number of features that can remain unidentifiable: some non-phonemic features must be identified now. This result is precisely what is desired, since, as was shown in the previous section, if a certain restriction on the number of unidentifiable features is not introduced at this point, it will be impossible to correctly apply the transformational rules of inflection and word production.
It should be noted that, according to the terminology of the Prague school, not fully identifiable morphonemes are analogues of "archiphonemes". Although Trubetskoy defined "archiphonemes" as "a set of meaningful features common to two phonemes", in his linguistic practice he operated with "archiphonemes", in which more than one feature was neutralized (unidentifiable); see his "Das mor- phonologische System der russischen Sprache".
Add to this that the rules of morphological structure provide for the application of transformational rules of Russian morphology to incompletely identifiable morphonemes, which are basically identical to the "archiphonemes" postulated by Trubetskoy in his work mentioned above.
1.57. The need to divide the F-rules into two groups and apply the MS-rules before transformations becomes even more obvious because in many languages there are significant differences between the restrictions imposed on sequences of segments within individual morphemes, and the restrictions imposed on sequences of segments. in general, without taking into account their division into morphemes. So, for example, in Russian, only very few vowel sequences are allowed within individual morphemes, while at the junctions of morphemes, any combination of two vowels is practically possible. In other words, in vowel combinations within morphemes, many features are non-phonemic and therefore must remain unidentifiable in the notation.
Many of the rules that identify these non-phonemic features can only be applied if individual morphemes are delimited from each other. However, during transformations, it is possible to rearrange characters in such a way that individual morphemes will no longer be delimited. The phenomenon of "syncretism" mentioned above can serve as an example of this. Another example is the so-called "interrupted morphemes", especially characteristic of the Semitic languages. "Interrupted morphemes" are also found in many Indo-European languages, including Russian. For example, in the neuter adjective (p’ust + o) “empty”, the sign “Neutral” is expressed by the fact that the stress falls on the stem and the ending (-fo). Since the morpheme delimitation may disappear during transformations, F-rules, which require information about the beginning and end of a morpheme for their application, must be applied before transformations.
1.58. After applying the rules of morphological structure, all segments appearing in the notation are either fully or incompletely identifiable morphonemes. Since morphonemes are uniquely defined by different paths on the tree representing the phonological system of the language, it becomes possible to replace the matrices with which various lexical morphemes are written with linear sequences of pluses and minuses, provided that a special symbol (in our case, an asterisk) will indicate the place where identification of incompletely identifiable morphonemes. No character is required to mark the end of identification of fully identifiable morphonemes, as this is determined automatically. In the following example, a space is introduced in such places to make them easier to read. However, unlike the asterisk, a space is a redundant character and cannot be included in the entry.
After applying the rules of morphological structure, the sentence presented in Table. I-1 can be written as follows:
The meaning of the + and - signs in this entry should be established using a tree depicting the phonological system of the Russian language (see Figure I-1). Pluses and minuses are commands that tell the tree to be scanned from top to bottom, always starting at the first branch point. In this case, the pluses indicate the need to select the right branch, and the minuses indicate the need to select the left branch. After selecting the terminal point of the tree or the point marked with an asterisk in the entry, the process starts again, from the first branch point. This procedure allows us to establish, for example, that the first segment of the entry above is an incompletely identifiable morphoneme, defined by the distinguishing features "non-vocal, consonant, non-compact, low tonality, tension."
1.581. An important consequence follows from the inclusion of incompletely identified morphonemes in the notation. Rao-
we look at the noun (*/*es) "forest" 81, in which in the plural and in the II-nd local case of the singular the stress falls on the case endings, and in all other cases of the singular - on the stem vowel. In the light of § 1.512, the genitive singular would be written as (*l'es+a) and the nominative plural as (*les+'a). However, since (*les+'a) and (l,is+'a) "fox" (as well as unstressed (e) and (1) in all cases) are homophones, it is necessary to add a rule that would contain the statement that unstressed ( e) goes into [i], or some other similar statement in terms of distinguishing features. However, in this way we include the unstressed (e) (and also the unstressed (o)) in the phonological system of the language, although these sets of distinguishing features are not used to distinguish statements. This is a direct violation of the condition (For-I), which specifically stipulates the impossibility of such a step. Since the condition (3a-1) was rejected by us as a requirement for phonological notation, such a violation is fully justified. However, it should be noted that there is an alternative to violating the condition. (Za-1), which consists in establishing several entries for all lexical morphemes containing the vowel sound (*e). So, for example, in this case (*l*es) would have to be written as /1,’es/ and /l,is-/, which undoubtedly complicates the notation in an undesirable way.
1. 6. Above, in § 1. 42, it was noted that after applying the transformation rules, including the rules of word production and inflection, the sentence record will consist only of phonological symbols, i.e. morphonemes and boundaries. The symbols of grammatical morphemes classes will be replaced by the phonological sequences that are derived from them, and the symbol # (vowel alternating with zero) will either be represented by a vowel or omitted from the notation. As a result, only the & symbol remains unidentified.
Condition (6): Symbols &, according to the rules of morphology, are transposed into phonological boundaries or excluded from the notation.
An exact description of the transposition process is part of the morphology of the language and therefore cannot be given in detail here. In this study, we will only list all types of boundaries and all the contexts in which they occur.
In Russian, there are five types of boundaries, which are indicated by the following symbols:
1) The boundary of a phonemic syntagm is indicated by a vertical bar |.
2) A word boundary is indicated by a space or, in cases where ambiguity may arise, by the % symbol.
3) The boundaries of prefixes and prepositions are indicated by the symbol =.
4) Before some endings, a special character 4 is placed, sometimes in the same cases, in order to avoid confusion, the symbol § is placed.
5) The boundaries of morphemes in abbreviations like (p'art-b, i * l'et) "party card" are indicated by the symbol - (dash).
Since the character & transposes only within the specified five kinds of phonological boundaries, all & characters that do not correspond to any of these kinds are eliminated from the notation. If in the course of the presentation it becomes necessary to somehow designate these junctions of morphemes, then for this purpose the sign (-) (hyphen) will be used, which, however, is not a symbol in the phonological notation.
1.7. Now we can continue to identify the sentence we have taken as an example. After applying the transformation rules of the language, we get the following notation:
This is the phonological notation of a sentence, since it only includes morphonemes and boundaries, and all the rules necessary to transpose this notation into sound describe only the effect of various configurations of distinguishing features and / or boundaries on individual complexes of distinguishing features.
Phonological rules can be formulated in such a way that there is no need to refer to the derivational history of morphonemes and boundaries. This requires the existence of a strict sequence in the application of the rules. If the rules are not ordered, their structure will become much more complicated, then it will be necessary to turn to the derivational history of symbols.
As an illustration, consider the following example. In Russian, all smooth and paired consonants soften before (*e). In addition, unstressed (e) becomes diffuse, i.e. [i]. The easiest way to state these facts is as follows.
Rule A: Before (*e), smooth and paired consonants soften.
Rule B: Unstressed (e) becomes diffuse.
However, if rule B is applied first, then rule A will need to be replaced by rule A":
Rule A": Before ('e[ and before [i], which comes from (e), smooth and paired non-compact consonants soften.
Obviously, rule A is simpler than rule A. However, rule A can only be applied when the order in which the rules are applied has been established.
Table I-2 shows the functioning of the phonological rules of the Russian language in relation to the sentence we have taken as an example.
At the initial stage, each morphoneme is written as a set of distinguishing features, which are interpreted using a tree (Fig. I-1), displaying
taken as an illustration (See 1.4 and
phonological structure of the Russian language. Further, after the application of individual phonological rules, morphonemes are modified. Since only a few P-rules are needed in our example, not all of these rules are shown in Table 1. I-2. The P 1 b rule is applied first, which assigns a voiced feature to morphonemes in which this feature is unidentifiable. Further rule R-2 applies. From the table, the functioning of this rule is clear. Subsequent rules are applied strictly in the order of their numbering until the list of rules is exhausted. As a result, we get the so-called "narrow" transcription of the sentence, which can be directly translated into sound:
3 1 2 3 3 1 4 2 1 4
| fcira | p,jani-jbrad,ag31Szokcerkaf, |
"Yesterday a drunken tramp burned down the church." The numbers above the vowel symbols indicate the degree of intensity of their pronunciation (dynamic amplification): 1- highest degree intensity, 4 - the lowest, the degree of intensity.
In principle, the phonological rules should apply until all the distinguishing features of all segments are identified, and these rules should also provide for the description of cases where a given feature is a free variant. Then it would be necessary, for example, to have a rule stating that all sonorants in Russian are always voiced (with rare exceptions like (o*kt, 'abr,*skoj) "October", where (r,) is often stunned). However, such rules are not included in the present specification. Since such facts are often also controversial, we decided that the value of such additional details would be very small.
2. Phonological system of the Russian language
When conducting a phonological analysis, the question always arises of the extent to which the proposed scheme of analysis takes into account the available data. In the description it is absolutely impossible to list all the phonological features of the speech of even one person, since he can use features characteristic of other dialects and even foreign languages (for example, a person speaking Russian can distinguish between nasal and non-nasal vowels in some (French) expressions, constituting an integral part of the colloquial vocabulary of this person). If we try to take into account such facts, it becomes obvious that a systematic phonological description is not feasible. Therefore, it seems appropriate to consider such cases as deviations and place them in special sections, and confine the bulk of the grammar to those facts that can be described systematically. This description considers a variant of the Russian language, basically identical to the variant described in such well-known works on the Russian language as the recently published academic "Grammar of the Russian Language" and the dictionary of Russian literary pronunciation edited by R. I. Avanesov and S. I. Ozhegov.
The so-called "literary" variant of the Russian language, described in these works, allows the existence of variants for some phonological features. In the present description, an attempt was made to take into account these options. It is interesting to note that such deviations do not affect the phonological notation of utterances, but the order and content of the phonological rules that transpose the phonological notation into sound.
2.1. Morphonemes. On fig. I-1 is a tree displaying the morphonemes of the Russian language. This scheme served as the basis for compiling the matrix of distinguishing features (Table I-3). The system includes 43 morphonemes; they are identified by 271 commands, each of which indicates the presence or absence of one or another distinguishing feature (+ or - in Table I-3 or branches in Figure I-1). Thus, it takes 6.3 commands to identify one morpheme. Condition (5) requires that the number of commands used in the record be minimal. In order to understand how
Rice. I-1. The scheme of the tree displaying the morphonemes of the Russian language. The numbers standing behind the branching ducks correspond to the following distinguishing features: 1. Vocality - non-vocality. 2. Consonance - non-consonance. 3. Diffuse - non-diffuse. 4. Compactness - non-compactness. 5. Low key - high key. 6. Tension - non-tension. 7. Nasality - non-nasality. 8. Continuity - discontinuity. 9. Sonority - deafness. 10. Softness - hardness. I. Impact - non-impact. The left branches correspond to minuses, the right branches correspond to pluses.
Since our circuit completely satisfies condition (5), we can compare the above figure with log 2 43 = 5.26 (5.26 is the lower limit achieved by reducing the number of instructions to a minimum). It must be emphasized that this comparison must be approached with caution: the sole purpose in this case is to show that the command reduction process led to very satisfactory results.
§ 134. Prague linguistic school was formed as a result of the activities of the Prague Linguistic Circle (1926-1953), headed by V. Matezius, and the participants - R. O. Yakobson, N. S. Trubetskoy, S. O. Kartsevsky, A. V. Isachenko, B. Trnka , B. Gavranek, J. Vahek, V. Skalichka and other linguists (mainly Russian and Czech). The phonological representations of the PLC participants are most fully and consistently presented in the book by N.S. Trubetskoy " Fundamentals of phonology"(published in German in 1939, Russian translation - 1960).
N.S. Trubetskoy, following F. de Saussure, consistently draws a distinction between language (general, permanent, existing in the minds of all members of society) and its specific implementation - speech (speech act). Since the sound units in language and speech are different, the two sciences of sounds are also different: phonetics studies the material side of sounds that form an unordered sound stream in speech using the methods of natural sciences, and phonology is their functional aspect (semantic distinction) in the language system by linguistic methods.
The most important thought of N.S. Trubetskoy, underlying his concept, is that all sound units - both speech (sounds) and language (phonemes) - have a feature structure, that is, they consist of sets of certain features. The task of the phonological description of a language is to single out among all possible features phonologically significant or differential signs (DP), i.e., those that participate in semantic differentiation. In this case, a phoneme is a language unit that has a unique set of differential features, that is, a set that does not coincide with the sets that characterize other phonemes.
§ 135. The beginning of a phonological description is the identification of meaningful sound oppositions of a given language. A sound opposition that can differentiate the meanings of two words is called phonological(semantic) opposition; examples of such oppositions are voice oppositions [t] - [d] ( tom house), place of formation [t] - [k] ( tom-com), method of formation [t] - [s] ( tom catfish), by hardness / softness [t] - [t "] ( toma-tyoma) In russian language; according to the length (tension) of vowels [i]- ( lead-lead) in English, etc. A sound opposition that cannot differentiate the meanings of two words is called an indistinguishable opposition; examples of such oppositions are cases of different pronunciation of [r] (front-lingual or velar) in the German word rate, different pronunciation [l "] (voiced or voiceless) in the Russian word dust, different [e] (more or less closed) in the Russian word this etc.: pronunciation [r] instead of [R], [l ^ "] instead of [l"], [e] instead of [e] does not change their meaning in these words.
Each member of the phonological opposition is a phonological (semantic) unit. The smallest phonological units are called phonemes. A phoneme is a set of phonologically significant (differential) features characteristic of a given sound formation. The totality of all signs characterizes sounds; only phonologically significant (DP) - phonemes. For example, in the words clatter, laughter, number, clatter, Christmas tree the first consonants have (among others) five features each (see Table 17); from a phonetic point of view, all these features are equal and define specific speech sounds.
Table 17. Articulatory signs of some consonant sounds of the Russian language
From a phonological point of view, the significance of these features is not the same. For example, the sound opposition of labialized and non-labialized consonants is never used in Russian to differentiate meanings, therefore the sign "labialized" is not differential in the Russian phonological system and does not characterize the consonant phonemes of the Russian language; in a position before a labialized vowel, the consonants are always labialized, but before a non-labialized vowel they are not.
All other signs can differentiate the meanings of words: the place of formation ( tom-com), method of formation ( tom catfish), deafness/voicedness ( tom house), hardness/softness ( toma-tyoma). Thus, [t] is opposed to [d], [s], [k] and [t "] in the same position before [o], which means that the signs are "deaf", "tooth", "explosive", " soft" are differential for him, and the phoneme /t/ is characterized by four DPs (this is the maximum set for consonant phonemes of the Russian language). At the same time, for example, [n] is included in meaningful oppositions only on three grounds: hardness / softness ( nose-carried), place ( legs-could) and the method of formation ( moose nose), and in terms of deafness/voicing there is no semantic opposition [н] // [н^] in the Russian language, therefore, the phoneme /н/ is characterized by only three DPs, and the feature "voicedness" is not essential for it ( integral). It should be borne in mind that not any insignificant feature of sound units is called integral, but only one that can be differential in a given language, but is not such for a given phoneme (for example, deafness / voicedness for Russians<ц>and
It is easy to see that /x/ in Russian has only two differential features (velar: ham-sam, slotted: move-cat), while /j/ has only one (palatal).
Table 18. Differential (in bold) and integral (in italics) phonological features of some Russian consonant phonemes
§ 136. The next stage of the phonological description is the identification of the composition of phonemes. N.S. Trubetskoy explicitly formulated the rules for identifying phonemes, which are still used (directly or indirectly) in all phonological theories:
If two acoustically and articulatory similar sounds do not occur in the same position, they are combinatorial variants of the same phoneme and are in relation additional distribution([i] and [s], [j] and [i9] in SLL, [r] and [l] in Korean).
If two sounds occur in the same position and replace each other without changing the meaning, they are optional variants of the same phoneme and are in relation free variation([zhur "i´] / [zh "ur" and´], [dust"] / [dust ^ "], [e´t't] / [e´t't]).
If two sounds occur in the same position and cannot replace each other without changing the meaning of the word, then they belong to different phonemes and are in relation phonological contrast (itch-court, house-dam-smoke-dum).
The phoneme can be realized by different sounds (for example, the phoneme /and/ can be realized by the sounds [and] and [s]). Any sounds in which a phoneme is realized, N.S. Trubetskoy calls variants of phonemes.
The phonological content of a phoneme is the totality of its phonologically significant features (common for all variants of a given phoneme), which is different from the totality of differential features of other phonemes. The phonological content of phonemes is determined by their entry into the system of oppositions of a given language, therefore, the phonemic composition is a correlate of the system of phonological oppositions.
§ 137. N.S. Trubetskoy developed a branched classification of oppositions on various grounds. We will consider only two such bases, the classification according to which is most widely used in modern linguistics.
Relationship between members of the opposition N.S. Trubetskoy highlights
private,
gradual (stepped) and
equivalent (equivalent) oppositions.
Oppositions are called privative, one member of which is characterized by the presence and the other by the absence of a trait (for example, aspirated / non-aspirated, labialized / non-labialized). A member of the opposition, which is characterized by the presence of a sign, is called marked, and a member of the opposition, in which the attribute is absent, is unmarked.
Gradual oppositions are called oppositions whose members are characterized by a different degree, or gradation, of the same feature - for example, low / middle / high vowels.
Oppositions are called equivalent, both of whose members are logically equal, that is, they are neither two steps of any feature, nor an affirmation or negation of a feature (for example,<п>//<т>: labial-dental).
According to the volume of semantic power or effectiveness in various positions N.S. Trubetskoy singles out
permanent and
neutralized oppositions.
Some oppositions are preserved in all conceivable contexts of a given language - they are called permanent (for example, the opposition<л>//<р>possible in SRLYA in all positions).
In other oppositions, the opposition of their members may not be carried out in all positions, in some positions it may be removed (neutralized) - such oppositions are called neutralizable. For example, the opposition of voiceless and voiced noisy consonants in SRLP is neutralized in the position of the end of the word - in this position, voiceless and voiced noisy consonants do not differ and, accordingly, cannot distinguish between words . Thus, in the position of the end of the word, the sign deafness/voicing is not differential, and the noisy consonants in this position are characterized by a smaller number of DPs than in the position before the vowel. Yes, in a word that the first consonant has a set of four DPs (plosive, dental, hard, deaf), since it is opposed to other sound units in all these ways (see § 135 above); the last consonant, which is in the position of neutralization according to DP deafness / voiced, has a reduced set of DP (explosive, dental, hard: cf. mole-cross, that-current, brother-brother), since this phonological unit does not have DP deafness/voicing - the sign "deaf" becomes integral in this position.
The unit of the phonological system, represented in the position of neutralization, is called the archphoneme . In the phonemic transcription of N.S. Trubetskoy, archiphonemes are indicated by capital letters corresponding to unmarked members of the opposition (for example, /toT/). An archiphoneme is a set of DPs common to two or more phonemes.. So, the phoneme /t/ (for example, in the word raft) has four differential features "tooth", "explosive", "hard", "deaf"; phoneme /d/ (for example, in the word fetus) - four differential features "dental", "explosive", "hard", "voiced"; at the same time the last consonant in the words raft, fruit, that has only three DPs - "dental", "explosive", "solid" (since in this position it is opposed to other phonetic units only on these grounds), and this set is common to the other two phonemes (/t/ and /d/ ). Therefore, the last consonant in the word that is not a phoneme (for a phoneme, the set of DPs is unique, not repeating), but an archphoneme /T/.
Table 19. Differential features of the phonemes /t/ and /d/ and the archiphoneme /T/.
An archphoneme representative can:
not coincide with any of the members of the opposition (for example, in English, voiceless aspirated and voiced non-aspirated consonants in positions after [s] are neutralized in voiceless non-aspirated),
coincide with one of the members of the opposition, despite the fact that the choice is due "from the outside", that is, the sound context (for example, assimilation); so, in Russian, deaf and voiced noisy are neutralized in the deaf in the position before the deaf and in the voiced in the position before the voiced noisy,
coincide with one of the members of the opposition, despite the fact that the choice is determined "from within" - i.e. does not depend on the properties of the position, but on the features of the systemic organization of the language; in this case, an unmarked member of the privative opposition acts in the position of neutralization (for example, in Russian, deaf and voiced noisy ones are neutralized in the deaf in a position before a pause) or an extreme member of the gradual (neutralization of vowels by rising in Russian in the first pre-stressed syllable after soft consonants is carried out in [and])
coincide with both members of the opposition (that is, in different positions, the archphoneme is represented by different members of the opposition). So, in German, the opposition [s] / [š] is neutralized in the position before the consonant at the beginning of the word in hissing ( Stadt), in other positions - in a whistling consonant ( best).
LITERATURE.
Trubetskoy N.S.. Fundamentals of phonology. M., 1960.
Trubetskoy N.S.. Morphonological system of the Russian language // Trubetskoy N.S. Selected works on philology. M., 1987.
Oliverius Z.F. Phonetics of the Russian language. Prague, 1974.
Prague linguistic circle. M., 1967.
Based on Saussure's division of "longue" and "parole", Trubetskoy N.S. creates his own phonological theory, based on the division of the science of sounds into phonology and phonetics: as a field of study of sounds from a physiological-acoustic point of view. Phonology, the subject of which is not sounds, but units of sound structure - phonemes. Phonetics refers to language as a system. Thus, phonetics and phonology, from the point of view of Trubetskoy, are two independent disciplines: the study of speech sounds is phonetics, and the study of sounds is phonology.
The only task of phonetics, according to Trubetskoy, is to answer the question: How is this or that sound pronounced?
Phonetics is the science of the material side (sounds) of human speech. And since, according to the author, these two sciences of sounds have different objects of study: specific speech acts in phonetics and the system of language in phonology, then different research methods should be applied to them. For the study of phonetics, it was proposed to use purely physical methods of the natural sciences, and for the study of phonology - proper linguistic methods.
When establishing the concept of a phoneme - the main phonological unit - N.S. Trubetskoy highlights its semantic function. Thus, the sounds that are the subject of the study of phonetics have a large number acoustic and articulatory features. But for the phonologist, most of the features are completely unimportant, since they do not function as distinguishing features of words. The phonologist must take into account only what, in the composition of sound, performs a certain function in the system of language. In his opinion, since sounds have the function of distinction and have significance, they should be considered as an organized system, which, in terms of the ordering of the structure, can be compared with the grammatical system.
From the point of view of the Prague School, phonemes are really unpronounceable. Being a scientific abstraction, phonemes are realized in various shades or variants that are pronounceable. But the phoneme itself, as an abstract unity of all shades, is really unpronounceable. Trubetskoy writes: the specific sounds heard in speech are rather only material symbols of phonemes ... Sounds are never phonemes themselves, since a phoneme cannot contain a single phonologically insignificant feature, which is actually not inevitable for a speech sound (Amirova T.A. , 2006).
The most comprehensive and systematic views of the representatives of the Prague School in the field of phonology are presented in the work of N.S. Trubetskoy "Fundamentals of Phonology", which is only the first part of the comprehensive work conceived by the author.
In 1921, Trubetskoy was the first in the history of Slavic studies to propose a periodization of the common Slavic proto-linguistic history, dividing it into four periods. To the first period, he attributed the era of the disintegration of the Indo-European proto-language and the separation of a certain group of “Proto-Slavic” dialects from among its dialects, explaining that “in this era, Proto-Slavic phenomena mostly spread to several other Indo-European dialects, especially often to Proto-Baltic, to which Proto-Slavic is closer Total. The second period can be characterized as an era of complete unity of the “common Slavic proto-language”, which was completely isolated from other descendants of Indo-European dialects, which did not have any common changes with these dialects and at the same time was devoid of dialectal differentiation. The third period should include the era of the beginning of dialect stratification, when, along with general phenomena, covering the entire Proto-Slavic language, local phenomena arose that spread only to individual groups of dialects, but they did not numerically prevail over general phenomena. In addition, during this period, the dialect groups themselves “have not yet managed to establish final strong ties with each other (for example, the West Slavic group as a whole does not yet exist, but instead of it there are two groups - Proto-Lussian-Lechitic, pulling to the east, and Proto-Czechoslovak, pulling south). The fourth period is the era of the end of dialect fragmentation, when general phenomena occur much less frequently than dialectical (dialect) phenomena, and groups of dialects turn out to be more durable and differentiated.
N.S. Trubetskoy was one of the first to substantiate the need for a tripartite approach to the comparative study of languages: the first - historical and genetic, the second - areal-historical (language unions, language zones), the third typological - and showed their application in a number of his works, among which stands out the final work on general phonological typology. In this area, in addition to many universals (they were later studied by J. Greenberg and other scientists), N.S. Trubetskoy revealed a number of more particular, local patterns. Thus, in the same article on the Mordovian and Russian systems of phonemes, he demonstrated an important phonological principle, according to which the similarity of the inventory of phonemes does not determine the similarity of their phonological functions and combinatorial possibilities. The latter in the Mordovian language are completely different than in Russian.
Although the interests of the young Trubetskoy lay in the plane of ethnography, folklore and comparison of the Uralic, "Arctic" and especially the North Caucasian languages. He, according to his autobiographical notes, nevertheless decided to choose Indo-European studies as the subject of university studies, since this is the only well-developed area of linguistics. After classes at the philosophical department and at the department of Western European literatures, where he stayed for a year (from the 1909/10 academic year), N. S. Trubetskoy studies at the then newly created department of comparative linguistics (primarily Sanskrit and Avestan).
At the same time, understanding phonology as “the doctrine of the sounds of a language, common and constant in the minds of its speakers”, and phonetics as the doctrine of the particular manifestation of the sounds of a language in speech, which has a one-act character.
Trubetskoy speaks of the relationship between both of these components of the doctrine, since without concrete speech acts there would be no language. He considers the speech act itself as establishing a link between Saussure's signifier and signifier.
Phonology is considered as a science that studies a signifier in a language, consisting of a certain number of elements, the essence of which is that they, differing from each other in sound manifestations, have a meaningful function. And also the question of what are the ratios of distinctive elements and by what rules they are combined into words, phrases, etc. Most of the features of the sound itself are not essential for the phonologist, since they do not function as semantic features. Those. it is the science of the language system underlying all speech acts.
Phonetics, on the other hand, considers physical, articulatory one-act phenomena. The methods of the natural sciences are more suitable for her. For her, the main questions are: How to pronounce the sound, what organs are involved in this. Those. it is the science of the material side of the sounds of human speech.
It should be noted that not all representatives of the Prague School of Linguistics shared exactly this opinion about the relationship between these two disciplines. N.B. Trnka believed that "the phonetician presupposes a language system and strives to study its individual actualization, while the phonologist investigates what is functional in individual speech and establishes elements that are determined by their relation to the whole language system." That is, thus, the main difference between phonology and phonetics for Trnka was the different direction of their research.
Returning to the solution of this problem in the Fundamentals of Phonology, it must be said that Trubetskoy defines three aspects in sound: “expression”, “address”, “message”. And only the third, representative, belongs to the sphere of phonology. It is divided into three parts, the subject matter of which is respectively: culminating language function (indicating how many units, i.e. words, phrases are contained in the sentence), delimitative function (indicating the boundary between two units: phrases, words, morphemes) and distinctive or meaningful, found in the explicative aspect of the language. Trubetskoy recognizes the semantic-distinctive function as the most important and necessary for phonology, assigning a special section to it.
Trubetskoy's main concept for semantic differentiation is the concept of opposition - opposition according to a semantic feature. Through the phonological opposition, the concept of a phonological unit (“a member of the phonological opposition”) is defined, which in turn is the basis for the definition of a phoneme (“the shortest phonological unit, the decomposition of which into shorter units is impossible from the point of view of a given language”).
As the main internal function of the phoneme, its semantic function is recognized. The word is understood as a structure identifiable by the listener and the speaker. The phoneme is a semantic feature of this structure. The meaning is revealed through the totality of these features corresponding to a given sound formation.
Trubetskoy introduces the concept of phoneme invariance. Those. the pronounced sound can be considered as one of the variants of the phoneme realization, because it, in addition to semantic differences, also contains signs that are not such. Thus, a phoneme can be realized in a number of different sound manifestations.
1) If in a language two sounds in the same position can replace each other, and the semantic function of the word remains unchanged, then these two sounds are variants of the same phoneme.
2) And, accordingly, vice versa, if the meaning of the word changes when the sounds are replaced in one position, then they are not variants of the same phoneme.
3) If two acoustically related sounds never occur in the same position, then they are combinatorial variants of the same phoneme.
4) If two acoustically related sounds never meet in the same position, but can follow each other as members of a sound combination. In a position where one of these sounds can occur without the other, they are not variants of the same phoneme.
Rules 3 and 4 regarding cases where sounds do not occur in the same position are related to the problem of identifying phonemes, i.e. to the question of reducing a number of mutually exclusive sounds into one invariant. Thus, a purely phonetic criterion is decisive here for assigning different sounds to one phoneme. Those. the interconnection of these sciences is manifested.
In order to establish the complete composition of the phonemes of a given language, it is necessary to distinguish not only a phoneme from phonetic variants, but also a phoneme from a combination of phonemes, i.e. whether a given segment of the sound stream is the realization of one or two phonemes (syntagmatic identification). Trubetskoy formulated the rules of monophonemic and polyphonemic. The first three are phonetic prerequisites for a monophonemic interpretation of the sound segment. A sound combination is monophonic if:
1) its main parts are not distributed over two syllables;
2) it is formed by means of one articulatory movement;
3) its duration does not exceed the duration of other phonemes of the given language.
The following describe the phonological conditions for the one-phoneme significance of sound combinations (potentially one-phoneme sound complexes are considered actually one-phoneme if they behave like simple phonemes, that is, they occur in positions that otherwise allow only single phonemes) and the multi-phoneme significance of a simple sound.
A very significant place in Trubetskoy's phonological system is occupied by his classification of oppositions. It was generally the first experience of this kind of classifications. The classification criteria for phonological compositions were:
1) their relation to the whole system of oppositions;
2) the relationship between members of the opposition;
3) the volume of their distinguishing ability.
According to the first criterion, the oppositions are divided, in turn, according to their "dimensionality" (qualitative criterion) and according to their occurrence (quantitative criterion).
According to the qualitative relation to the entire system of oppositions, phonological oppositions are divided into one-dimensional (if the set of features inherent in both members of the opposition is no longer inherent in any other member of the system) and multidimensional (if the “grounds for comparing” the two members of the opposition extend to other members of the same system) . Oppositions are quantitatively divided into isolated ones (the members of the opposition are in relation to those that are no longer found in any other opposition) and proportional (the relationship between members is identical to the relationship between members of another or other oppositions).
On relations between members of the opposition:
Privat oppositions: one member of the opposition is characterized by the presence, and the other by the absence of a feature: [e] - [n] - everything is the same, except for nasality.
Gradual - the sign is graduated: the degree of rise in vowels.
Equivalent (equivalent), where each of the members of the opposition is endowed with an independent feature: [p] - [w] - one labial-labial, the other - labial-tooth.
Permanent and neutralized oppositions: [deaf] - [voiced] in Russian - a neutralized opposition (the phenomenon of stunning - voicing), and in German and English these oppositions are permanent.
As a special section of the “phonology of the word”, the Prague Linguistic School singles out morphonology, the object of study of which is the phonological structure of morphemes, as well as combinatorial sound modifications that morphemes undergo in morphemic combinations, and sound alternations that perform a morphemic function.Along with the synchronic description of phonemes, the Praguers tried to define the foundations of diachronic phonology, based on the principles:
1) no phoneme change can be accepted without recourse to the system;
2) every change in the phonological system is purposeful.
Thus, de Saussure's thesis about the insurmountable barriers between synchrony and diachrony was refuted.
Phonology (from Greek phōnē - sound + logos - word, doctrine) is a branch of linguistics that studies the sound side of the language in its functional significance, in other words, the theory of phonemes.
The central place in phonology is occupied by the doctrine of the phoneme as the shortest (indivisible in time) unit of the sound side of the language, which has a distinctive (distinctive, semantic) ability (crowbar, com, rum, volume, catfish, etc.).
General phonology deals with the analysis of the essence of the phoneme, elucidating the relationship between the phoneme as a sound unit and the sounds representing the phoneme in the flow of speech, on the one hand, and between the phoneme and morpheme, phoneme and word, on the other. It establishes the principles and methods (rules) for determining the composition (inventory) of the phonemes of a language, as well as the oppositions in which they are located, and the connections that exist between individual phonemes or their groups, which make up a single system of phonemes - a phonological, or phonemic, system. .
The scope of the concept of "phonology" in different linguistic schools is defined differently.
However, any of them deals with the variability of the phoneme, establishes a system of phonemes and their modifications.
Phonology originated in Russia in the 70s. 19th century Its founder was I.A. Baudouin de Courtenay, who introduced the concept of "phoneme" (a unit of language), contrasting it with the concept of "sound" (a unit of speech).
The successor of the ideas of the scientist of the late period was his student L.V. Shcherba, who in 1912 revealed the sound factors that determine the division of speech into phonemes, and pointed to the semantic-distinctive function of the phoneme.
Initial ideas of I.A. Baudouin de Courtenay developed N.F. Yakovlev, who made an important contribution to the development of phonology in the early 1920s. 20th century
Based on the ideas of these scientists, phonology received further development and world recognition in the works of the Prague Linguistic Circle.
Major phonological schools
Kazan Linguistic School. Representatives: I.A. Baudouin de Courtenay,N.V. Krushevsky, V.A. Bogoroditsky and etc.
KLSh is a linguistic direction of the 20th century, the main provisions of which are:
- building a theory of phonemes: 1st understanding of the phoneme, i.e. the phoneme was understood as a generalized type, as a mobile element of the morpheme; 2nd understanding of the phoneme, i.e. the definition of a phoneme as a mental representation of a sound;
- phonetic alternations due to the strict distinction between evolutionary and statistical language learning;
- the allocation of phonetic units (coherents and divergents on the one hand, and correlatives and correspondents on the other), which cannot be identified with sounds; attributing the issue of phoneme alternation to the “alternation theory” and, on the basis of historical alternations, creating new science- morphonology - and the introduction of the concept of "morphoneme".
Moscow phonological school. Representatives: R.I. Avanesov, P.S. Kuznetsov,A.A. Reformatsky, V.N. Sidorov, A.M. Sukhotin, N.F. Yakovlev, M.V. Panov and etc.
IPF is a linguistic direction of the 20th century, characteristic features which were:
- 1st understanding A.I. Baudouin de Courtenay formed the basis of the Moscow phonological school, as a result, the phoneme is determined by the morpheme, the phoneme is understood as a series of positionally alternating sounds that may not have any common phonetic features;
- the starting point in the views on the phoneme is the morpheme, i.e. the phoneme was determined through the morpheme: the identity of the morpheme determines the boundaries and scope of the concept of the phoneme, and the sounds of weak positions are combined into one phoneme not by their acoustic similarity, but by their functioning as part of the morpheme (in the words valY and voly, unstressed vowels, despite the identity of the sound, represent different phonemes, because in the first case there is a phoneme<а>(cf. shaft), and in the second - the positional version of the phoneme<о>(cf. ox); the final consonants in the words fruit and raft represent different phonemes, because in the first case, the positionally transformed phoneme<д>(cf. fruits), and in the second - a phoneme<т>(cf. rafts));
- the definition of two main functions of phonemes: perceptual (the ability of a phoneme to identify) and significative (that is, the ability of a phoneme to distinguish morphemes);
- differentiation of types of alternation - crossing (options) and parallel (variations) types;
- theoretical development of the concepts of "neutralization" and "hyperphoneme";
- consider soft [g '], [k '], [x '] variations<г>, <к>, <х>, as well as [s] - a variation<и>;
- distinction between a phoneme in a narrow sense, the so-called strong phoneme, formed by the main type of a phoneme and its variations - members of a parallel (non-intersecting) alternation, and a phoneme in a broad sense, the so-called phonemic series, a set of sounds formed by the main type of a phoneme and its variations - members of a non-parallel (intersecting) alternation, i.e. identified strong and weak positions of phonemes.
St. Petersburg (Leningrad) phonological school. Representatives: L.V. Shcherba, M.I. Matusevich, L.R. Zinder, L.V. Bondarko and etc.
SPFS is a linguistic direction of the 20th century, the main provisions of which were:
- 2nd understanding of the phoneme A.I. Baudouin de Courtenay formed the basis of this school, where the phoneme is defined as a historically formed sound type that serves to distinguish between words, to create words that are potentially related in meaning;
- when defining the concept of phoneme, they proceed from the word form, in which, according to the physiological and acoustic feature, the shortest sound units are distinguished (in the word wol, the vowel phoneme is<о>, in another form of the same word oxen in an unstressed position, in accordance with the sound, a phoneme is distinguished<а>; in word forms fruits and rafts, consonant phonemes in the final syllable -<д>and<т>, but in the original form of these words, fruit and raft have the same final consonant phoneme<т>);
- believe that soft [g '], [k '], [x '] are not variations<г>, <к>, <х>and [s] - not a variation<и>, but the actual phonemes;
- the absence of the concepts of phonetic variants and variations (the phoneme is close to the sound and is determined by the sound in speech), however, there is a separation of the shades of the phoneme - combinatorial and positional.
Prague Linguistic School(Prague linguistic circle, school of functional linguistics). Representatives: Czechs - V. Mathesius, B. Gavranek, B. Palek, B. Trnka,J. Vahek, V. Skalichka, Russians - N.S. Trubetskoy, S.O. Kartsevsky, R.O. Jacobson, FrenchA.Martine and etc.
PLS is a structural-functional direction in linguistics of the 20-40s. XX century, creatively combined interest in the internal correlation of language units, their semiological nature with attention to their extralinguistic functions and connections with extralinguistic reality (examines language in connection with common history people and their culture).
Achievements of PLS in the field of phonology, in particular in his work "Fundamentals of Phonology" N.S. Trubetskoy singled out the provisions:
- distinguished between phonetics and phonology on the basis of units of speech (phonetics) and language (phonology);
- defined the phoneme as a scientific abstraction realized in its pronunciation variants: "a set of phonologically significant features characteristic of a given sound formation";
- substantiated the concept of a phonological system, outlined the main sound functions: culminative (vertex-forming), delimitative (delimiting), distinctive (sense-distinguishing);
- in phonemes, he singled out distinctive (differential) features that make up the content of phonemes;
- singled out the method of opposition (lat. oppositio - opposition) as one of the leading ones in the field of studying the properties of the phoneme;
- developed a system of oppositions.
Thus, the differential features of phonemes are manifested using the opposition method:
a) relations between members of the opposition:
- privative - when one member of the opposition has a sign, while the other does not:<в>and<ф>, <д>and<т>;
- gradual - one sign can manifest itself to a greater or lesser extent (longitude and brevity of sound):<ā>and<ă>, <д:>and<д>;
- equipotent - the members have a completely different set of features: and<ц>, <п>and<р>.
Members of the opposition form a correlative pair.
b) base according to the volume of semantic power:
- constant - in a certain environment, phonemes retain their characteristics:,<у>,<н>;
- neutralizable - in a certain environment, phonemes lose their signs and retain only common features: <б>, <э>, <з>.
These oppositions are transferred to other levels of language learning. It is important to study the syntax.
