When I say “All communication is manipulation,” it is sometimes protested that many utterances are advanced without a clear goal in mind.
The first issue here is that one need not hold a desired outcome consciously in mind in order for one’s behavior to be shaped by, attracted to, and in the service of a desired outcome. We are not transparent to others or ourselves.
But I will readily grant that many communicative moves do not have specific outcomes in mind (otherwise we would be constantly disappointed). Rather, utterances can be strategically aimed with varying levels of specificity or narrowness. I will carve these into four classes of goals: precise, vague, generic, and bespoke.
Precisely strategic utterances are uttered with a specific end result in mind—for instance, telling a police officer a lie in order to avoid a traffic ticket. There is a single, binary outcome kept in mind which is either met or not met.
Vaguely strategic utterances attempt to shift an interaction or situation toward a vaguely better situation, or away from a vaguely worse situation—for instance, avoiding certain topics in professional settings to avoid unspecified blowback, conflict, or disharmony. One steers with, and follows, the winds, instead of seeking a singular destination.
Finally, generically strategic utterances are those which are uttered for a large audience, attempting to achieve a number of vague or specific goals among the largest possible swathe of audience—for instance, presenting a consumer product as organic in order to increase its buyer appeal among certain demographics, or presenting a specific political front with the precise goal of winning an election. This is in contrast to what we might consider “bespoke” utterances.
Of course, generic and bespoke, vague and precise, lie on a spectrum. No utterance is crafted fully generically or totally bespoke. No desired outcome is perfectly specific or totally vague. In other words, manipulation is always value-neutral on certain axes. We may specifically wish to avoid a traffic ticket, but be agnostic as to whether the officer, during the traffic stop, makes small talk about the weather, his golden retriever, or the evolved origins of back pain. And even when we believe that our interactive desires are at their most open-ended, that we are perfectly agnostic to whatever response we might receive, we can easily find ourselves surprised, upset, enraged, or disappointed by the emotional tenor or semantic content of the response.
Sometimes we say, “This carving restrains us; it is limiting or oppressive; it constrains what we are able to say or think.” Or we might say, “This carving is insufficient, it gets us into many problems.” This is true in a sense, but it is perhaps more accurate to say, “This carving was generically useful for a number of historical functions, and marked an improvement over whatever conflation preceded it historically. But as the circumstances have changed, or as I myself have sought from these terms a specific utility differing from that which the carving is fitted to, the terms have proven insufficient, or downright counter-productive.” In this frame, it is less about attributing absolute, context-less failure to a carving—since few distinctions survive the generations if they are not useful in some way—and more about attributing a misfit between a carving, an environment, and a task at hand.
The “fit” of a word to a task can in some sense be understood through statistical distributions. We might usefully distinguish between gender, referring to a mode of cultural expression, and sex, referring to a biological state. (Of course, because of current controversies, these distinctions are messy and under continual renovation—but grant them here.) This distinction serves a purpose because there is a statistical discrepancy between biological identity and cultural presentation—there are many members, or a sizeable set, whose biological identity differs from their cultural presentation. But in a society, perhaps a historic society, where the two were functionally coupled, there would be no need for separate terms; the two would co-occur.
As an example of such a case, the Jahai, a Malay Peninsula rainforest community, have lived continually in heavily forested areas. As a result, they have never needed for a distinction between the concept of “outside”/”outdoors,” on the one hand, and the concept of forest on the other. Both are merely hüp. Among the Avatime of southeastern Ghana, the central distinction is between cultivated farmland and uncultivated wilderness, liŋwàfù. This wilderness is defined as being the location tribe members go to create new fields for cultivation—or occasionally, to hunt. While there is a modified concept, liŋwàfù tsìtsì, for wilderness which cannot be cultivated—more permanent wilderness—the distinctions are not based on the type of vegetation, e.g. the separation between large brush and trees. Most modern Western languages have an enormous number of possible landscapes—brush, tundra, forest, swamp, meadow, shrubland, etc. “Plant communities” can be defined by their species, floristically, or phytophysiognomically, by physical structure. Habitats can be defined by climate (polar, temperate, tropical, Mediterranean) or vegetation type (forest, steppe, grassland, desert). These are commonly known because a modern Westerner must navigate a social and informational environment in which these distinctions matter—he must understand that “Los Angeles is in the desert”—that it is a climate in perpetual drought, which needs outside reservoirs—or that his drive through the Mojave will be low on scenery—or merely that he might like the desert, and wish to live there one day.
Similarly, when we think of “species,” we think of both relatedness and mutual ancestry. This is in large part because, while convergences do frequently occur, the holistic relatedness of two same-species organisms practically speaking is always the result of shared ancestry. Were we to discover nearly identical genetic species on another planet, who had no shared ancestry with our own species, we might need, suddenly, to make a decision about how to arbitrate such distinctions—we might refer to species in the relatedness sense as X-species, and species in the ancestral sense as Y-species, to account for this statistical discrepancy in sets. What was previously unnecessary, because the features were coupled, has become necessary now that they are uncoupled (and because, ostensibly, there is some functional or pragmatic advantage in distinguishing these types). The same is true of any attempt at distinction, for instance, in selection games, assessing parties use surrogates, or heuristics, to distinguish between qualified and unqualified candidates. But these heuristics by definition are fitted to, and thus base their effectiveness on, a specific state of the environment. One famous case is the changing meaning of fair vs. tan skin, in historical vs. contemporary Western society.
This is also why the philosophical counterexamples beloved by 20th C analysts are confused at their core. The carvings we make are the result of specific, concrete, statistical distributions of attributes and features, in relation to specific, concrete, statistical distributions of tasks and purposes, references and distinctions. To imagine a counterexample on a “Twin Earth,” with water that is like water in every way except molecularly, is to misunderstand language and meaning: what it is, how it comes into being, how it is continually renovated and made to work by humans for human tasks. First, it is impossible and incoherent to suggest that this twin water would be like our own H2O in every functional way, but be structurally different; every structural difference will make a pragmatic difference given the right goal or game. Further, were humanity to have evolved, and evolved language, on a planet with both types of water, we would colloquially call both “water” (since they fulfill the same functional purposes) in all settings that did not require a distinction between types, and likely make a distinction only in those technical cases where a functional difference surfaced.
We can see clear cases in which an instrumental purpose of distinction drives the differentiation of subtle varieties, the splitting up of one word or concept into many—for instance, the need for more complex and precise counting systems in cultures with complex economic systems. Agricultural or “cultivator” cultures have more folk categories for plants than do hunter-gatherers, even controlling for environment (Berlin 1992, Balee 1999, Voeks 2007). There is very little reason for a pre-industrial society to distinguish between one-hundred thousand, a billion, and the concept of infinity—these amounts are functionally uncountable; they are all “countless.”
In short, there is no universal, objective taxonomic method, and no universal objective answer as to what, taxonomically (and thus linguistically) speaking, things “are” or “aren’t.” (Nor is there such a method for distinguishing what is one thing, or a separate thing, or two things entirely.) Our classification and thus our reference schemas are pragmatic, and pragmatic taxonomies rely not just on our purposes and goals (what we wish to do with our distinctions, be they heuristics in our selection games, or descriptive categories in our language) but also our environment and its distribution of difference. We are making distinctions between sets, in order to fulfill desires. As our purposes, and the environment (including but not limited to the distribution of assessed objects) changes, so too does the optimal taxonomy. Taxonomy is defined by fit between goals and environment. In that sense, it is a tool for winning games, where a game is defined as the situation of an agent, embedded in an environment, with preferential states of being that drive goals.
 See also intrinsic empowerment.