Occasionally a word creates two or more incompatible interpretations within what appears to be a single context. Example:
``I could not care less.''
1. I could not care any less than I already do.
[That is, I already don't care at all.]
2. I could not possibly care less than I do now. [I am already caring at my maximum. In other words, it is not possible for me to care any more than I already do.]
3. I could not care less [at the moment, but I might care more in the future, OR I might continue to trust at my current level...]
`Care Less' (vs Careless!)
[This reminds us that the English language actually has 27 content-bearing literals, with the space being essential to distinguish `to get her' from `together' and `care less' from `careless'.]
Then, what about removing the `not' in the above:
``I could care less.''
This also has similar ambiguities.
So, it seems that `less' causes some difficulties, which brings me to the bone of my contention!
Recently I heard the term `trustless', which caused me to pause and reflect on what it might mean.
What might we mean by the following statement?
``X is trustless.''
This seems to be inherently ambiguous, with two conflicting interpretations:
1. X cannot be trusted.
2. X does not trust others.
Part of the problem is the missing notion: who is trusting whom? Indeed, `trust' is not an absolute concept, and generally needs to be qualified -- who is trusting whom, with respect to what assumptions, under what logical conditions, temporal aspects, physical locations, and other caveats.
Assume each of X and Y is a person, computer program, data item network, cloud, or whatever you like, where for the moment X is in some sense distinct from Y. For example, we consider the following cases (an oversimplification that avoids the middle ground between case 1 and case 3, and between case 2 and case 4 -- involving sets of Ys).
1. X cannot be trusted: No Y should trust X. (ForAll Y, default)
2. X cannot trust others: X does not trust any Y. (ForAll Y, default)
3. X cannot be trusted by some particular Y. (ThereExists some Y)
4. X does not trust some particular Y. (ThereExists some Y) The issues of conditional quantification could considerably complicate the possible interpretations of `trustless'.
However, harkening back to Mike Schroeder's thesis on mutual suspicion, there could be asymmetry in mutual trustlessness. For example, system X might trust user Y only if several authentication algorithms were successfully executed on X; user Y might trust system X only if several integrity mechanisms were invoked on X. More generally, we could necessarily have many different forms of mutual trustlessness.
The above discussion involves the question of the direction in which the `trustless' relation goes:
Cases 1 and 3: Y is trustless with respect to perceiving
less trustworthiness in X,
Cases 2 and 4: X is trustless with respect to perceiving less trusthworthiness in Y.
So, what about reflexive trustlessness? Perhaps you might think that the potential direction of the ambiguity does not matter so much when X is self-trustless. But the ambiguity is still there.
X is trustless with respect to X (under specified conditions).
For example, I might not trust myself when I am drunk. Here the directionality of trustlessness might seem to be somewhat clearer. But am I trustless with respect to my drunken self? Or is my drunken self trustless, because it is not to be trusted? That is, is it my sober self that is trustless, or my drunken self?
The bottom line is that `trustless' is a very wobbly term, and should probably be avoided, even if the writer attempts to define it very precisely and only locally. Furthermore, we must distinguish carefully between `trust' and `trustworthiness' (that is, `worthy of being trusted').