The Name-Leash Hypothesis: Maternal Separation and the Origins of Displaced Reference

Displaced reference — the ability to refer to what is not in the perceptual field — is a defining feature of human language, and the usual account of its origin has it emerging from calls that recruit others to distant food. That account has a weak point: where pointing or leading would serve, an offline representational system is not obviously worth its cost. This paper proposes a different first case. In the cooperative-breeding arrangement reconstructed for early Homo, a mother handed her infant to an allomother and left to forage, so that mother and infant were placed, day after day, in perceptual situations that no longer specified one another. The problem this poses is not distance but partition: how to go on holding one specific, non-fungible individual once one's own situation supplies no index of it. A perceptually grounded gesture has nothing to work on there. An individualized vocalization does — and the paper argues that a proper name, rather than a food call, is the more plausible candidate for the first displaced signal. The semiotic core is the broken index. An index signifies through a connection that holds at the moment of signifying; when the referent goes, that connection lapses and cannot be restored by producing the signal again. What takes over its office is a sign of another type: a vocal form tied to one individual, keeping the singularity of pointing while working without perceptual contiguity. The paper separates four stages that the label "proper name" otherwise runs together — an infant-associated form, its displaced use by the mother alone, a shared mapping organized by the referent rather than by any caller, and a public name — and places Kripke's rigid designation at the last of these rather than at the first. Around this core the paper assembles evidence from cooperative breeding and life history, upright walking and vocal-respiratory control, developmental phenomena such as solitary babbling and the fort/da game, comparative work on vocal individuation in dolphins, chimpanzees and marmosets, and neural findings on cross-modal identity representation. It distinguishes three levels of decoupling — from arousal, from here, and from now — and locates the proposed mechanism at the third, which the published comparative record does not document. A Claim–Evidence Map in the accompanying Supplementary Materials gives the status of every claim the mechanism makes: observation, established background, inference, reconstruction, or prediction. The hypothesis is offered as a reconstruction, not a demonstration. The conjunction of conditions it describes is put forward as sufficient for the route it reconstructs, not as necessary for any route to displaced reference. Changes in version 4 Version 4 does not revise the hypothesis. It changes how the hypothesis is stated, and it calibrates the evidence behind each step. Composition. A new §1.2 states the mechanism in one pass, as a self-contained account a reader can take entire before meeting any evidence; the introduction is reorganized around it (§§1.1–1.3), and everything that follows examines one link of that chain. The name trajectory. §3.3 separates four stages — infant-associated form, displaced individualizing use, shared referent-centred mapping, public proper name — and states what each requires. Rigid designation in Kripke's sense is now presented as the profile of the fourth stage rather than as the step that opens the first. Displacement and conventionalization are treated as separable: the first displaced use may well have been private. Time. §2 replaces delay and distance with a partition between the signaler's situation and the referent's, together with stability of the sign across changes of situation. The criterion is presented as an operational bridge to the problem rather than as a definition of it, and the paper is explicit that nothing in it requires an interval to be given in advance. Two invariances. The identification of an individual as the same across encounters and modalities is distinguished from the shared referential invariance that circulation of a form produces. The first is presupposed by the mechanism and neurally provided; the second is its product. The callitrichid comparison. The contrast with marmosets is restated as two working adaptations to distributed infant care rather than as a deficit on one side, and a short statement of the resulting mismatch — between transfer ecology and attachment architecture — is added to the main text (§8.5). Comparative material. §8.3 is substantially expanded around the recent work on vocal labelling in marmosets and the discussion it has provoked. The comparative table now marks reconstruction as reconstruction and absence of evidence as absence of evidence, with a note on how to read it. Evidential calibration. §9 is divided into plausibility anchors and proposed implementations, and says which is which. Throughout, reconstructions are no longer stated in the grammar of findings. Figure. Figure 1 is redrawn to match the four-stage account. Supplementary Materials. Two sections are new: a Claim–Evidence Map (S0) giving the status of every claim made in §1.2, and a treatment of alterocentric reference and the distributed present (S6). The bibliography is now divided: this document's reference list covers sources cited here, and sources cited only in the Supplementary Materials are listed there.