It turns out that a pile of documents on the floor near the door to an attorney's office is rather like the message beep on your crackberry, or a peacock's feathers: all are signals that convey qualities of the signaler.

Usually, signaling theory is applied to the analysis of communication, and lately to the analysis of informational artifacts in social media [1, 2, 3]. In these analyses, a perceived signal - whether it is the presence of large plumes on a peacock, the display of jewels at a social occasion, or the number of friends one has on a social network profile, are all a means for the signaler to communicate the possession of a characteristic to the receiver. Consequently, signaling theory defines a signal as a communication that intentional, and beneficial to the signaler to produce, and is beneficent. These analyses tend to focus on the information asymmetries of the signaler-receiver relationship (reliability & honesty), as well as attempting to discover the various kinds of signals extant.

All of this assumes that the signaler is not the receiver. In the case of studies of computer mediated communication, this additionally assumes that receiver and sender are human. What if either of these assumptions are not valid?

Let us consider the first case - the signaler is the receiver - in the light of the phenomenon of piling on the floor. During a recent project on work processes in law firms, we saw a lot of people leaving documents out, in plain sight - on the floor, under the desk, even leaving cabinets and drawers open so the contents are visible. Of course, these are classic cases of situated cognition, or (some may claim) distributed cognition - the person making things visible is doing so as a strategy to remember to do something about it: a case of the world as memory.

What is interesting about this strategy is that it is essentially an act of delayed communication to oneself. This is especially true when things are placed so as to get in the way - at the threshold to the office, for example, or at a corner of one's desk. However, it is not a perfect strategy: people forget anyway. One explanation of this is that the signal is unreliable - it no longer provokes an action, or triggers a recollection. This is because it is a conventional signal: it does not cost the signaler much (other than acquiring a reputation for messiness) to produce the signal. Too many piles on the floor sitting around for too long, and the signal becomes meaningless. Of course, that is not the only contributing factor - this strategy is more useful for a repeatable, predictable work process than a complex, fragmented one in which several strategies might be competing for attention, and hence for scheduling work, and it is also better suited to short-term reminders, such that the pile on the floor does not fade into the background because of being present and unmoving for so long.

Which brings us to the second case - a non-human communicant, and, of special interest to designers of human-computer interaction. The increasing ubiquity of portable computing devices raises questions of their social impact, and the consequences of constantly interacting with something. Part of this is a search for calmth, and on another level, an exploration of spirituality. But the underlying question is this: how should information technology in general, and personal mobile devices in particular, be present in the user's life?

Mobile devices can be irritating. The constant beeping of a blackberry is similar to Windows Vista's much hated security prompts - the communication takes much the same form for issues/items that have a range of complexity or importance.

Consider the computer as signaler, and the person as receiver. Communications from the computer - message boxes, alerts, beeps, vibrations are all signals that the user's attention is required. However, the fact of requesting attention is not sufficient for the user to make an informed decision about whether the attention is merited - because the signal is essentially binary, and has only two codes: on and off. As a result, people develop strategies for coping with constant interruption - ignoring the alerts, treating them all the same, bypassing them, or just turning the damned things off.

One way of fixing this problem is just to increase the richness of the signal - use more codes. Different sounds for emails from close colleagues, and silent notifications for your monthly linkedin update, for instance. Other ways are to move the signal to the periphery of attention, make the signal context-sensitive (this usually means time & location), or - and this is a cultural resolution - reduce the need to use the devices in the first place.

However, there is a more interesting - and perhaps more perverse - option. To return to signaling theory: conventional signals are kept reliable not by the cost of production, but by the cost of having the deception discovered. In human society, these are usually social costs, with consequences like exclusion or expulsion, incarceration, and the loss of trustworthiness or reputation [5]. In other words, the cost of signaling is paid by the device, when it is dishonest - when it fails to request attention appropriately. Now this leads to some rather exciting scenarios - "slapping" my blackberry for beeping when I'm trying to write this blog post, for instance - but while the prospect of perpetrating violence on one's computers is quite compelling, it is quite disruptive. It might give me great pleasure to bang my cellphone against the table if it goes off during a meeting, but it might not make me very popular with my colleagues. I might also not have the time or attention to devote to socialising my phone - I'll probably just turn it off altogether, as a defensive strategy.

Not all penalties are expensive to enact: getting ignored when attempting to pick someone up at a bar, or a raised voice by an angry parent are often sufficient penalties and relatively easy to inflict. Of course, they are only effective because the cost is perceived by the offender, and future costs can be predicted based on the penalty. Blackberries have no such model of social interaction.

So this is the design challenge: should we design computers (and especially mobile devices) as signalers with a predictive cost model, and if so, how? Should we make phones that are 'happy' when the owner responds to an interaction request as predicted, and 'sad' when he/she doesn't, like the elevators in the Hitch-Hiker's Guide to the Galaxy (this is often confused with giving computers personality and affective expressivity, which have a host of other interactional uses as well).

From a Star Trek movie: Scotty steps into an elevator of the newly (and badly) re-designed Enterprise after some frustrating mechanical problems on the bridge...

Scotty: Engine room

Elevator: Thank you

Scotty: And up yours...

References

1. Donath, J., & Boyd, D. (2004). Public Displays of Connection. BT Technology Journal, 22(4), 71-82. doi: 10.1023/B:BTTJ.0000047585.06264.cc
2. Donath, J. (274). Signals in Social Supernets. Journal of Computer-Mediated Communication, 13(1), 231-251. doi: 10.1111/j.1083-6101.2007.00394.x
3. Lampe, C. A., Ellison, N., & Steinfield, C. (2007). A familiar face(book): profile elements as signals in an online social network. In Proceedings of the SIGCHI conference on Human factors in computing systems (pp. 435-444). San Jose, California, USA: ACM. doi: 10.1145/1240624.1240695
4. Yates, J. (2007, June 9). CrackBerrys: Exploring the Social Implications of Wireless Email Devices. MIT Sloan School of Management. Retrieved April 22, 2008, from http://mitworld.mit.edu/video/471/
5. Scott-Phillips, T. C. On the Correct Application of Animal Signalling Theory to Human Communication. In A. D. M. Smith, K. Smith, & R. F. I. Cancho (Eds.), The Evolution of Language: Proceedings of the 7th International Conference on the Evolution of Language (pp. 275-282). Singapore: World Scientific.