In the situation of interlingual contrasts the well-known principle of asymmetric dualism of the linguistic sign, which was formulated, in its monolinguistic variant, by Serge Karcevskij [1], appears as asymmetry in the division of the semantic continuum and the repertoire of formal means in various languages. A certain correlation between languages can only be found at the universal level, in the form of a potential of formal means and functional grammatical integrals, which unite elementary meanings (atomic senses or proto-functions). In each particular language both expression and content are divided into discrete units in a specific manner, or, it would be more careful to say, a specific way of correlating expression and content 'units' from a universal inventory of possibilities is chosen.

Let us recollect one principle used in practical translation. It is the principle of non-linear translation, when the 'word-after-word' procedure is avoided. What is actually translated is not a separate word (form), but a bigger block of the utterance, i.e. context is taken into account. Contextual means contribute to reproducing the original grammatical meaning. In practical translation a separate word (form) can not be regarded as a unit of correlating texts, in contrastive studies it is also more appropriate to look for a more extensive unit of comparison than a separate word (form). When there is no parallel form, or when an existing parallel form is functionally insufficient to convey the required content from the original language, the original meaning is dismembered, and the 'atomized' units are re-distributed between contextual and lexical (and sometimes, formal) means in the target language. In the process of translation, meaning does not appear as expressed 'globally' or 'synthetically', i.e. pertaining to one grammatical form. It is expressed discretely, 'analytically', i.e. within a grammatical-contextual complex [2], which can unite heterogeneous means. Re-quoting Bulgakov, one can say that meanings, like manuscripts, don't burn. As they do not disappear in interlingual transitions, this can be regarded as the law of conservation for linguistics.

Grammatical units, like phonemes, include atomic constituents of non-equivalent value. Some of these elementary meanings are in oppositional relationship, some constitute the material context for this opposition. Among these elementary meanings, one can find those which are in latent opposition. Being latent, as well as transition from latent to overt oppositional relationship are treated here from the point of view of the universal human language, and a possibility of latent oppositions becoming actual in some particular language is taken into account. A very similar idea can be found with S.D.Katznelson who pointed out the fact that a grammatical form, besides categorical functions, expresses a whole functional complex. He also noted that overt, or outer grammar, which finds its embodiment in grammatical forms of this or that language, is based upon covert, or inner grammar [3]. In some sense, one can find, in any language of the world, and in the shape of underformed, non-grammatized, potential oppositions, atomic (elementary) meanings, which appear as grammatized in some other language. This is actually what can be understood under covert (latent) grammar in one of the senses of this term [4]. It is possible that these hypothetical elementary meanings (senses) are in a certain hierarchical, systemic relationship towards each other. Anyhow, as compared to the periodic system of chemical elements, the search for these universal elementary senses and their hierarchy seems to be several degrees higher in difficulty and accessibility for the researcher.

It is not out of pure scholastic sophistication that differentiation between opposition proper and the material base for this opposition is suggested. In the situation of language contrast, it is the latent context of the opposition that brings up the majority of problems. Overt oppositions can sufficiently quickly and relatively successfully be acquired by language students, even if in their mother tongues they do not appear as overt. A phonematic example can be furnished: French [é/è] phonemes present a certain difficulty for Russian students of French. This difficulty is often removed by quoting two Russian words (not phonemes!) эти/этот [eti/etot], in which a positional variant 'brings to the surface' a latent opposition of open/closed. Latent sphere is the source for the possibility of learning proper, and in some cases it displays its quality as the potential for a possible self-development of the system. Another example: English accent in the Russian о can be attributed to the non-differential – for an English speaker – material elements in the Russian triphthongoid [уОа] [uOa], etc.

Grammatical integrals taken as wholes, as well as types of grammatical-contextual complexes, are not just chaotic sets of occasional senses. The fact that many languages display repeated combinations of such elementary meanings, such complexes of atomic senses, and even the fact of mutual translatability itself brings up the idea that grammatical integrals and functional types play the role of attractors in the presumably chaotic process of human linguistic (grammatical) activity, thus displaying an example of collaboration between accident and necessity [5].

A grammatical form within a, say, 'perfect' ('article', etc.) language presents a basis for unity in functional diversity. In interlingual contrasts, if one of the languages is 'perfect-devoid' ('article-devoid', etc.), the relevant functional types retreat to the potential, covert domain. For instance, the semantic spheres of ‹definiteness/indefiniteness›, or of the ‹present perfect› never disappear in Russian, or in any other language where these functional semantic fields are of the so-called non-focused type. Formally, a semantic zone, or the functional potential of a universal grammatical integral, finds its representation in formal means that belong to different levels of language structure, but get united in a complex. For example, the ‹present perfect› is expressed in a wide range of types within a continuum with a common semantic denominator [6]:

I have lived here since my childhood	Ya zhivu zdes' s detstva.
I have lived here all my life.	Ya prozhil zdes' vs'u zhizn'.
I have never lived there for such a long period.	Ya nikogda n'e zhil (n'e zhival, n'e byval) tam tak dolgo.
When I have lived here long enough, I’ll know for sure.	Kogda ya prozhivu zdes' dostatochno dolgo, ya budu znat' nav'ern'aka; etc.

U men'a (typologically equal to “ya imeyu” = I have) uzhe vs'o sd'elano.		I have done everything.
Gd'e on u teb'a tam spr'atan? (Dostoyevsky)//	where have you put/hidden it?// German: wo hast du verwahrt?// French: Où l’as tu mis/caché e?// Ital.:. dove l’hai nascosto?//

Ya eto v Parizhe slyshal, ot odnogo frantsuza, chto... (Dostoyevsky)//

I heard it in Paris from a Frenchman// German: Ich hörte es in Paris von einem Franzosen// French: J’ai entendu dire à Paris par un français// Ital.: Un francese, a Parigi, mi disse// Hung.: Párisban egy francia hozta fel, azzal a megjegyzéssel, hogy...// Bulg.: Slushal sum go v Parizh ot edin frantsuzin, che... //

One of the vivid examples of transition from contextual collaboration of means to the stage of grammatized form is presented in the development of the so-called analytical forms. The proto-grammatical function is first assumed by the lexical meaning of the auxiliary component syntagmatically, then, paradigmatically: the re-semanticized auxiliary component is grammatized. The representation of the universal integral ‹progressive› in Danish can serve as an example. Some elementary meanings which are connected with this integral, are expressed through verbs of motion: Thomas går og driller, literally: goes and laughs (= is laughing); Peter sidder og ser fjernsyn, literally: sits and watches (is watching) [7].