Let’s suppose four worlds (W1, W2, W3, W4).
World 1 contains mothers (x), maternal aunts (r), and fathers (z), but not shamans (k):
W1 = (x = x) ∧ (r = r) ∧ (z = z) ∧ ∄k
World. 2 also contains mothers (x) and aunts (r), but (as it happens, e.g., among the Iroquois; I am drawing this example from Patrice Maniglier) they are both called mothers (x); and it contains fathers (z), like World 1; but, in contrast World 1, it also contains shamans (k):
W2 = (x? = x + r) ∧ (z = z) ∧ (k = k)
World no. 3 contains mothers (x) and maternal aunts (r) distinguishable from one another, as in World no. 1; it contains fathers (z), as World 1 and World 2 do; and it contains shamans, like World 2, but in this case two entirely different types of shamans (as it happens, e.g., among the Mekeo), namely, a war shaman (p) and a peace shaman (q):
W3 = (x = x) ∧ (r= r) ∧ (z = z) ∧ (k? = p) ∧ (k? = q)
Lastly, World 4 contains mothers (x) and maternal aunts (y) distinguishable from one another other, shamans (k), and fathers but unlike in the preceding worlds, the fathers) don’t play any significant household role (as it is the case, e.g., among the Navajo).
W4 = (x = x) ∧ (r = r) ∧ (z? = j) ∧ (k = k)
Truth (what is to be a mother? a maternal aunt?, a father? a shaman?) can only be seen here as variation; put otherwise: insofar as falls always-already somewhere in between and across all such worlds, it must be deemed translatable rather than universal – for there is no model-world, no paradigm (W0) – or else particular.
Variant no. 1: truth never equals a natural number, but an infinitesimal number.
Variant no. 2: truth is always approximate.
Corollary: If sameness were to be granted primacy over difference (if being a mother, a maternal aunt, a father or a shaman were to mean one and the same thing), translation would be superfluous. Conversely, if difference were to be granted primacy over sameness, if difference were to be declared incommensurable), translation would be impossible.
(a) Leibniz’s philosophy, where certainty is always approximate in between the unfolding of two infinites; (b) Lévi-Strauss’s study of myth, beautifully epitomized in the formula: fx(a) : fy(b) :: fx(b) : f1/a (y); and (c) Guattari’s schizoanalytic meta-modeling, where four invariable abstract functors (F, Φ, U, T) may take any imaginable empirical values… they variously supply the best theoretical references I can think of for a redefinition of truth as – let’s put it plainly – interval.
Yet translating is how we all permanently live – what I manage to know falls somewhere between what I thought I knew and what I will yet get to know, what I manage to build falls always between my memories and my expectations, and love is also the experience of an ever-enriching in between…
Postscriptum: If you still think there is some universal truth like, say, 2 + 2 = 4, please consider that the Daribi (as Roy Wagner remarks) “only” have three numbers: “one” (deri), “two” (si), and “many” (sera) – which is quite sophisticated indeed, as for a series of the type: n + 1 = x can also be written as: n + 1 = y – 1, a third term being in either case, here too, that which helps to make sense of the series; which means that Daribi mathematics is serial in the last instance and in its own way. Accordingly, it would be possible to say that 2 + 2 = many, which is not exactly the same as saying that 2 + 2 = 4, but is not entirely different either. That is to say, “4” and “many” are, in this specific context, truths transversal to each another, transversality being obviously different from universality.
