More New Math
Writer and artist Craig Damrauer has another series of New Math to share. Some are better than others, but I like the simple concept and execution overall.
Blogless
One of the big problems we isolated last week with our consequentialist strategies was that they couldn’t seem to handle the problem of justice. Since we know that design needs ethics, and since ethics ought to be able to deliver justice, we suggested on Wednesday that consequentialism (evaluating designs in terms of their consequences) — or at least the varieties of it we’ve undertaken so far — may be the wrong approach.
The alternative approach we suggested is that a design could count is good if it instantiated or promoted some agreed-upon design virtues. I personally find something intuitively compelling in the idea that justice is a virtue of design, and thus that one necessary (although almost certainly insufficient) condition for a design to be a good design is that it exhibits, expresses, or promotes justice. I don’t expect it to be contentious that justice is desirable, so rather than argue directly that design should concern itself with justice, I’ll explain why a few advantages of the idea of trying to develop an ethical code around justice as a design virtue.
I am sure you expected that like our consequentialist solutions, an ethical code that mandates instantiating or promoting a set of design virtues (a virtue ethical code) has some distinctive advantages and some distinctive disadvantages.
On the one hand, it has the enormous practical advantage of being able to provide a code of ethics that matches up with our intuitive sense of what a good design might be (e.g., it should be just, and so on).
A particularly convenient facet of this is that virtue ethics comes packaged with the advantage of making our motivations count for something. If a design is good that promotes (e.g.) justice, there is a certain register on which the same design cannot be bad, regardless of its consequences, if the designer merely tried to make a just work. After all, when designers try to act in a just manner, this is in itself a form of promoting the virtue of justice inside the practice of design. So, apparently, virtue ethics might come with a more satisfying algorithm to determine blameworthiness than consequentialism.
This is due in part to an ambiguity in the idea of applying virtue ethics to design. Philosophical virtue ethics typically concern themselves with the inner states of individuals - an action counts as good because the agent who brings it about was motivated by a virtuous motivation. The analog of this is for design is the idea that a(n object of) design would count as good if the designer made her design choices in a virtuous way. I think that there is a perfectly reasonable concern about the applicability of this ethical model to design for the precise reason that designs and actions have very different ontological statuses. I’ll pick this thread up on Wednesday.
Finally, the notion of instantiating virtues is one to which I think many designers may naturally gravitate. I have no specific facts or arguments for this, except for the career choices of our various paragons of design ethics. For example, I think that Tibor Kalman must have subscribed to something like virtue ethics to make the choices he did.
While this last bit may not initially strike the consequentialist as particularly persuasive, it is worth considering that among those designers who produced the best consequences, many may have been trying to instantiate or promote virtues — and that is something that the consequentialist will have to take seriously.
The Stockholm-based design firm BVD is advertising their new line of products, available from the Japanese office supply company Askul. Among them is this clock, which I find worthy of reflection.
Consider the following: This is a beautiful clock, whose form follows its function. It is as easy to use in the context of an office as it is clever and visually striking. It is, by these counts, very nicely designed indeed. Things get less clear, though, when you consider that its core function seems to be reminding you every time you look at it that you’re somewhere you don’t want to be.
There’s a great deal of talk these days about a desire for transparency –transparency in our software, our banking system, our government, etc.– but transparency itself may not be enough.
Case in point: your standard legal contract is transparent. The complete documentation is made available to all parties involved. The dilemma is that the average person can’t understand it as it is full of complex legal language and procedure. Granted, contracts are constructed this way to hold up to scrutiny, but I think it begs the question: if only lawyers can understand the document, can we really call that transparent?
Aviary, a collection of browser-based design tools, summarizes its Terms of Service in plain English. In doing so, I’ll bet more users will actually read them (I did).
It’s a lesson, I think, that transparency alone doesn’t always fulfill its intended function. While policies are important, we should be careful not to forget the role of user experience.
Transparency must be designed.
After all, maybe consequences aren’t everything. We’ve certainly seen that they’re difficult to quantify, and that makes it difficult to come up with an algorithm that produces intuitive and edifying results. What’s worse, we’re often wrong or only partially right when we try to predict the consequences of our actions.
We’ve also seen that for all forms of consequentialism, one’s obligations to oneself and to one’s family and friends are no weaker or stronger than those to strangers, a fact which seems to skew certain types of situations towards having to choose between an intuitive, non-maximally ethical conclusion or a non-intuitive, maximally ethical one.
It’s also worth noting that consequentialism seems to make the somewhat confusing assumption that blameworthiness is not really a moral quality internal to an agent, but rather a function of whether or not it is societally useful to blame her. In the case of the designer who takes the job designing cigarette packaging because she needs money for her sick mother, we don’t necessarily blame the agent because we hold her as particularly blameworthy in a moral sense, but rather because it is useful for us to blame her if we want to advocate against products that endanger the health of children.
Which is all to note that consequentialism might be called an agent-neutral theory. Starting next week, we’ll consider an alternative theory, one that is agent-based. For now, let’s call this an agent-based virtue ethics.
In our first, very rough formulation of this, we might say that a good design counts as good by virtue of the fact that it is undertaken with virtuous motivation on behalf of the designer, as well as exhibits, expresses, or promote certain design virtues.
Does this idea, which goes back to ancient Greece, help us intuitively solve the kinds of moral dilemmas we’ve been posing so far? That’s something to think about until Monday, when we’ll discuss it more.
The imperfections in Nickel’s vintage-print style are as charming as his designs. It’s a refreshing aesthetic in our often all too perfect world of computer illustration.
Consider the following scenarios:
These scenarios are meant to highlight a problem with mixed-criteria maximizing.
Consider the terms that our hopelessly beleaguered designer might consider in her decision whether to take the engagement in the first scenario:
The point is that there is a very reasonable possibility that (a)+(d) > (b)*(c). However, I imagine that most of us would intuitively balk at the idea of taking this engagement, despite the fact that the consequentialist scales balance in favor of our doing just that. The reason for this, it strikes me, is that maximizing just doesn’t have a provision built in for distributing utility evenly. Rather, it maximizes brutely.
Let’s hold that thought and look at our second scenario.
In the first scenario, we ran up against a situation in which, despite the fact that a certain outcome seems to be maximally preferential, it still feels wrong to bring it about. The possibility of encountering this problem is brought out even more sharply by our second scenario.
Consider the following features of our second scenario:
Despite this, most of us would do almost anything that was within our power to save a loved one’s life. Almost certainly we would attempt to rationalize our decision (”If I don’t do it, someone else will, so I might as well,” etc.), but at the end of the day, the consequentialist scales will simply never balance in favor of the personal project. Regardless, I am sure, many of us would take up the engagement.
And few would blame us for doing so. Indeed — as we hinted at above, and articulated explicitly in our discussion of satisficing (or mixed-criteria maximizing) — a practical consequentialist algorithm for evaluating the rightness of an action intuitively puts choices in preference order rather than in an impartial order of maximal utility. Since this is the case, apparently if a personal project is important enough to us, a satisficing algorithm can deliver us any result we want. This means that algorithms of this type, despite delivering highly intuitive results, tend to be rather unedifying.
This, combined with the fact that neither form of consequentialism seems able to vouchsafe the equal distribution of utility, implies that neither is particularly well-equipped to handle the problem of justice. This strikes me as a relatively significant handicap. I’ll introduce a possible alternative on Wednesday.
