The study of totalitarian manipulative practices forms within a system of social sciences a specific research field that has large intersections with linguistic pragmatics, theory of speech influence and political linguistics as well as with social and political psychology and theory of argumentation, to name only the disciplines that are broad enough by their scope. It, however, neither coincides strictly nor can be treated as a proper subsection of any of them (except, perhaps, political linguistics whose limits and internal structure is not clear yet). Being relatively young, the study of totalitarian manipulative practices, nevertheless, possesses a certain tradition, uses some field-peculiar notions and rests upon certain assumptions which, to my mind, deserves critical consideration.

1. REFLECTIONS ON MANIPULATION. The central notion of the field, that of manipulation, is by no means a neutral scientific term. Definitely it is an evaluative word that, as a matter of fact,  has a manipulative potential of its own. Indeed, the word manipulation is used by an external observer Z to characterise approximately the following situation:
a) there is an actor X  who, pursuing a certain goal Gx, effectively makes another actor Y behave so as to promote Gx; and
b) Y does not realise that his/her behaviour is influenced  by X (or at least does not care about this); and
c) Z, having in mind an ideal model of rational behaviour and some idea about Y’s “true interests” Iy, believes that  Gx and Iy may not be (and, in prototypical case, is not) coincide, and therefore condemns X.
In most cases the notion of manipulation is narrowed by stipulating that “effective making” is achieved by communicative means.

In other words, the semantic structure of the natural language word manipulation, used as a term, contains three main components: technical (effectiveness), cognitive (unconsciousness), and moral/social/political (negative evaluation). This structure corresponds to the main incentives to study manipulation (see below), but what is also very important is that a study of manipulation becomes, by virtue of terminology used, “conviction-based”. And moral/social/political condemnation depends upon one’s one moral/social/political beliefs and is an instrument for imposing such beliefs. Calling someone a manipulator means admitting his/her communicative effectiveness plus condemnation of his/her communicative activity; to put it another way, the relation in the pair communicator [President Reagan was called a Great Communicator, as we remember] vs. manipulator is the same as in freedom fighter vs. terrorist or, to use E. Herman’s example, faith ‘my deeply held beliefs’ vs. fanaticism ‘his deeply held beliefs’, etc.
Totalitarian regimes of the 20th century have been universally condemned, therefore the study of their communicative technologies (to use a more euphemistic term) or “linguistic crimes” (to use overtly evaluative wording) is a standard and respectable research practice; but modern Western-style regimes have also been repeatedly accused in political manipulation by their liberal (and not only liberal) critics (to put aside what has been said about advertising, PR, etc.). Moreover, in modern Russia the term “manipulation” (also “robotisation”, “zombiing”)  is used mostly by opponents of post-Soviet reality, and respective practices are attributed to instigators of Perestroika in the late 1990s (it is enough to say that the most cited title now is “Manipulation of Concsiousness”, a huge volume  by Serguei Kara-Murza, a principled and inventive defendant of “the Soviet Civilisation”).

2. THE ROLE OF LINGUISTICS AND SEMIOTICS. Even seen from purely technical stance, that of effectiveness, manipulation is not an entirely linguistic phenomenon. Numerous handbooks of “covert influence” tend to treat manipulation in terms of complex discrete “techniques” in which not only linguistic, but also semiotic, psychological (affective and cognitive), logical and social mechanisms are amalgamated. But, however, to analyse manipulative discursive practices one should be able to isolate linguistic aspects of manipulation from all others. In particular, I would like to dwell upon the semio-pragmatic basis of manipulation, which, I believe, is the readiness of a human to ignore slight, but semantically relevant differences between linguistic forms as being irrelevant. It is commonly believed that nearly endless variation of expression gives language its power and flexibility; but the ability to ignore difference makes communication cognitively feasible – and, in the same time, gives rise to language misuse when communicants’ selectivity differs considerably. The low selectivity of the addressee my be conditioned by satisfaction and lack of real interest in analysing more thoroughly the incoming flow of verbal information or by lack of the ground for comparison (exclusion of such a ground is a function of censorship), or by both.

3. THE INCENTIVES. As I have already stated elsewhere,  there are three main motivations behind research projects in political linguistics: moral/social/political discontent (“Orwellian motivation”),  applied tasks (an attempt to obtain knowledge about mental properties of communicators from supposedly objective text-analytic procedures), and technical curiosity  (a desire to test linguistic analytical instruments against an interesting class of texts). Now I am in the position also to state a strict correlation between the first and the last motivations, on the one hand, and respective aspects of the semantic structure of the word manipulation on the other, as well as a loose correlation between the middle members.

4. A DESCRIPTIVE PROGRAMME. A research landscape in both the study of totalitarian manipulative practices and political linguistics reminds those of theoretical linguistics of the early 19th century: strong generalisations then were put forward on a rather shaky empirical grounds. Orwellian Newspeak plays for the field virtually the same role as Latin and Greek played for linguistic theory before the development of comparative (first historical and then typological) studies. What has been said about Newspeak tends to be taken for granted as a universal truth about all kinds of totalitarian discursive practices, which it probably is not.
In fact, what is being studied in political linguistics are, first and foremost,  different idiopolitical discourses. A special interest to totalitarian species, besides moral reasons, is heated by their more prominent and more obvious peculiarities (just like formal literary studies tends to be concentrated on modernist experiments). Different types of political discourses should be investigated comparatively to make generalisations more valid.

5. COMPETING TREATMENTS OF LINGUA SOVIETICA. A comparative study of idiopolitical discourses might also contribute to the isolation of their truly linguistic characteristics. Generalisation about totalitarian discourses are not only often premature (why not, after all; some premature generalisation of linguistic typology are still alive); what is more serious is that they are basically non-linguistic, or, to put it less categorically, non-structural. The existing interpretations of the Soviet kind of totalitarian discourse known as Lingua Sovietica (I do not speak about a complete ignorance of its historical evolution and internal heterogeneity here) borrow their basic assumptions from logical positivism (Orwellian idea of Newspeak as the language of lie and its development in conceptions of “Wooden language”, etc.), from anthropology (a ritualistic conception), from sociolinguistics (to be more precise, from the theory of language situations; I mean the conception of diglossia as applied to the language situation on the Soviet Union) and from cultural and religious studies (a conception of  esoteric communication). All these interpretation are surely interesting, but external and therefore, as can bee shown, incompatible which each other (e.g., Orwellian conception presupposes a conventional relations between plane of expression and plane of content, while in true diglossia it is not the case, etc.). Given the complex and not entirely linguistic nature of the notion of manipulation the reliance upon non-linguistic conceptual instruments might well turn inevitable, but the admittance of this must be preceded by more intensive explorations in purely linguistic characteristics of totalitarian discourses. Some of them have already been successfully revealed (parallel nominations, deagentivised syntax, overuse of first person plural, emotional overloading), but how delimitative and exhaustive this peculiarities are remains to be seen.