I study dictionaries. The English language contains an archaeological stratigraphy representing great historical shifts in the composition of society and the gradual evolution of meaning and usage. Words do not stand still, and any attempt to fix them will fail. Some words change fast; compare Jane Austen’s ‘economist’—an efficient lady of the house—with today’s academic fraternity, others less so. Perhaps trust belongs to the latter category.
The Germanic register of English stresses the active voice of ich; it is personal and transitive. Latin‑based constructions yield agency to abstractions, camouflaging the speaker’s voice in passive or intransitive verbs. Beyond this ancient distinction, modern society has recently exposed us to public and private bureaucracies that have removed from sight the human interdependence that once gave rise to the words wuere now use. Contemporary standard English frequently collapses the distinction between subjective persons and mental or material objects, so that trust can refer equally to a person, idea, or thing—a shopkeeper, the national government or a type of machine. Nothing stands between the individual subject and a depersonalized, abstract world; and our words reflect this. But in origin they referred to another world explicitly founded on concrete human relationships made actions and sensations: coercion, persuasion and falling in love.
The rediscovery of such a world was largely an achievement of British social anthropology in the middle of the last century. By investigating stateless societies in Africa and Oceania, my predecessors revealed patterns of social organization and human interdependence that we had forgotten and could not recreate in our abstract speculations. This lent power, for example, to African Political Systems (Fortes and Evans-Pritchard 1940) and their monographs on the Tallensi and the Nuer. Unfortunately, these lessons are easily lost. We need this knowledge if we are to escape from the exclusive tyranny of modern conceptions of the world. For this reason, I will distinguish carefully between trust in persons and more abstract extensions of the idea, so that we can compare what we are with what we are not.
The set of synonyms to which trust belongs is unusually confusing.[1] Faith, trust, and confidence all express belief. This was originally something personally held dear—beloved. To believe is to accept something as true, to have faith, trust, or confidence in someone or something. Faith and confidence both come from Latin fides, which is the nearest thing to Germanic trust. Since the middle ages, faith has been the strongest word for belief in a collective idea or leader. To cut through the details, I propose a continuum of words connoting ‘belief’ based on how much they rest on the evidence of the senses. Faith is immune to contrary evidence and experience; trust is an expectation based on inconclusive evidence, b*6eing tolerant of uncertainty or risk; confidence is a strong conviction based on substantial evidence or logical deduction. These are all subjective attitudes.
The notion of reliance expresses complete confidence, a presumptively objective state where belief is no longer necessary. To rely on something is to be tied to it, bound as to an objective condition of existence such as religion. When one side is dominant, reliance becomes dependence. Belief, by contrast, is en extension of past experience into a future that is unknowable. It is a feeling that a person, idea or thing will not fail in performance. Feeling varies inversely with evidence or proof. Thus, Faith is an emotionally charged, unquestioning acceptance. Trust implies depth offeeling as reliable, with inconclusive evidence or proof. Confidence involves less intensity of feeling, being based often on repeated past evidence for being sure of regularity. Trust thus stands in the middle of a continuum of words for belief mixing extremes of blind faith and open‑eyed confidence. Its etymology shows trust to be true like a tree, firm, steadfast, and loyal; not impervious to the evidence of the senses, but founded on a willingness to endure risk and uncertainty.
Trust has been historically associated with the notion of friend. Modern usage restricts the number of our friends to a narrow circle of intimates, narrower by far than those we might occasionally be inclined to trust. But among the Frafra migrants—and for Englishmen of earlier times—trust was bound up with an idiom of classificatory friendship which had a relatively broad application. The American Heritage Dictionary—the most recent systematic compilation—defines ‘friend’ as ‘a person whom one knows, likes, and trusts’. This is better than the Victorian Oxford English Dictionary’s retention of Dr. Johnson’s definition: ‘one joined to another in mutual benevolence and intimacy’, if only for its fewer syllables. A frien is someone to whom one is not bound, and hence etymologically speaking the relationship is free, based on choice not obligatory status. A kinsman is rarely a friend. Friendship is free and analogous to wife in some Germanic languages—Frigg, Friya, wife of Odin, someone who chooses and is chosen. If friendship no longer means to us what it once did, nor does trust when we can place trust in persons, ideas and things with no semantic discontinuity.
Social relations have generally evolved from a concrete, interpersonal basis to the mass anonymity of urban crowds, an evolution obscured by a modern usage that tends to objectify persons as abstractions, much as preindustrial cultures personify objects. Social theorists perennially ask how qualities of trust, fairness, and co-operation that belong naturally to local worlds might be projected onto the modern stage of world markets and nation-states. Mainstream economists, however, by basing their most general propositions on a formal rational calculus of individual decision‑making, obscure the distinction between personal and impersonal dimensions of social experience on which this essay hinges.
I agree with the pragmatic emphasis in the original volume (Gambetta 1988) when both Partha Dasgupta and Niklas Luhmann insist on the importance of personal agency in defining trust. Dasgupta identifies trust as an expectation about another person’s actions which influences our own actions before we have full knowledge of the outcome; and Luhmann exposes to full view the distinction between personal and impersonal social organization. Clearly trust is at base the predication of one’s own actions on the actions of others bearing an identifiable risk of turning out unfavorably. The problem is to control extrapolation from rational choice approaches to the smallest and largest frameworks of modern mass society.
The institutions of industrial society may sometimes justify a position where the personal‑impersonal distinction remains blurred. But the present study of life on the margins of the modern state requires us to keep the difference between levels of social experience in sharp focus. A similar problem arose when anthropologists turned to the analysis of kinship in stateless societies. Here the largest political groups, such as clans, are often represented as being like families. Yet Meyer Fortes (for example, 1958) showed that the political and domestic levels were often contradictory and should be kept analytically separate.
An anthropological approach to trust would insist on no less a degree of analytical discrimination. To summarize, I situate the notion in a set of belief concepts bounded at the extremes by faith and confidence, where the variable significance of evidence or proof is matched by a corresponding level of emotion. As such, trust is in the no-man’s land between status and contract, the poles of primitive and modern society in evolutionary theory (Maine 1861).
[1] The principal sources for the etymological argument are the appendix on Indo-European roots in the American Heritage Dictionary (Morris 1969: 1505‑50) and Buck’s Dictionary of Selected Synonyms in the Principal Indo‑European Languages (1949). Much of this work is highly speculative, even poetic.*