Product Design
The Mach Effect
A lot of the core visual elements of the Marginalia branding exploit an optical illusion called the Mach effect. If you look at the concentric bands of color on the labels, or the adjacent bars and boxes of color on the website and logo, you will probably notice that there appears to be a gradient in each band or stripe, so that each band of color seems to get lighter as it approaches a darker color. This is an optical illusion. The pigment or digital color setting is completely uniform in each individual band or block, and your visual system itself fabricates the appearance of a gradient within each band of color.
To get a clearer, more visceral sense that this gradient is illusory, study your visual experience of the bands on the edges rather than in the middle. You’ll probably see that the effect is pronounced in the second band, but almost absent in the first. Then cover the first band entirely and watch as the effect you had seen on the second band diminishes. I have no expertise in the psychology of perception, but my sense is that our attention is drawn to things that load the perceptual system with extra processing tasks. The Marginalia labels and logo are, therefore, straightforward attempts to manipulate you into paying attention to the wine. If you’re reading this, maybe it worked?
Bottles
I would love to use clear bottles to show off the colors and clarity/turbidity level of the wine, but I decided to use darker glass because it offers protection against light damage, which can occur in as little as half a hour in clear glass that is exposed to bright light, or especially sunlight. Over the next few years I’ll be experimenting with lighter color glass (never clear though) and lighter weight packaging, but for now I think the wine is best protected in a darker, mid-weight bottle. In the absence of more transparent packaging, I also hope that the labels give some impression of what sort of experience you might find in the bottle, by supporting the stylistic/descriptive names of the wines: red wine, light red wine, and amber wine.
Corks
First of all, I am committed to avoiding natural corks. Despite best efforts, cork taint has not been convincingly addressed by the natural cork industry. There are very expensive corks that are guaranteed to be taint-free (free from trichloroanisole (TCA) and other haloanisoles), but apart from this, they offer no better or more consistent performance than mid-range corks. The product I’ve used since the 2013 vintage for most of my wines is the DIAM closure. This is a technical closure made mostly from natural cork, but ground up, processed to remove all TCA, and reformed into an agglomerated cork with specific oxygen transfer rates corresponding to different products (I use the DIAM 5, which has a lower transfer rate than a 3, but higher than a 10, etc.). In my estimation, the DIAM closure does a better job of consistently protecting the wine than any natural cork. It has a lower long term oxygen transfer rate than most natural corks, and doesn’t introduce nearly as much oxygen as a result of compression and insertion in comparison with natural corks. Because sulfites are the main weapon against oxygen, all of this supports my commitment to keeping sulfite levels as low as possible.
The design on the side panel of the corks has nothing at all to do with wine, though the concentric shapes echo the visual vocabulary of the labels and logo. In the other side of my professional life I have spent a lot of time thinking about the foundations of mathematics. The main logical/mathematical theory we use to study the foundations of mathematics is set theory. Set theory details the logical structure of a very minimal notion of grouping, but if elaborated in sufficient detail, it can characterize the logical structure of nearly any branch of mathematics. It is a cumbersome language to be sure, but it is also the closest we’ve come to a universal language for all of mathematics. While set theory can describe sets of any sort of objects you like, in the mathematical and philosophical study of set theory, we often want to consider the universe of pure sets, that is the universe of sets that can be constructed from nothing whatsoever.
There are a number of ways of describing this universe and its features, but students usually meet it first as V, the von Neumann hierarchy, which begins with the empty set and builds up in stages by collecting all subsets of the previous stage.
If we represent sets by drawing brackets around the things in the set, then we can draw the empty set as an empty box.
[ ]
This is a collection with no members, which sounds like an odd thing to think about, but if we want to study the pure logic of this mode of collection, without mixing in any accidental features of the things we collect, then it makes sense to consider an empty set as a limiting case.
Let’s call the empty set V0.
To build the stages from this point, we define the notion of a powerset: the powerset of a set X is the set of all subsets of X. A set Y is a subset of Z if and only if all members of Y are also members of Z. So, for example,
[2,5,4] is a subset of [1,2,3,4,5],
because each member of the former is also in the latter, but
[5,4,7] is not a subset of [1,2,3,4,5],
because the former contains 7, and the latter does not.
Since there is nothing in the empty set at all, there is noting in the empty set that is not also in any given set. Therefore, the empty set is a subset of every set, even itself.
And this is how we get something from noting: The set of all subsets of the empty set is, it turns out, not empty, since it contains the empty set (and noting else). Let’s call this V1:
[ [ ] ]
Continuing in a similar vein, V2 is the set of all subsets of V1:
[ [ ], [ [ ] ] ]
We can pick up a pattern when we look at V3, the set of all subsets of V2:
[ [ ], [ [ ] ], [ [ [ ] ] ], [ [ ] [ [ ] ] ] ]
Here’s a way of drawing attention to the salient details:
V0 has 0 elements
V1 has 1 element = 2^0
V2 has 2 elements = 2^1
V3 has 4 elements = 2^2
The pattern seems to be that Vn has a number of elements equal to 2 raised to the power of the number of elements in the previous stage, Vn-1. Once you get into the swing of things, it isn’t too too hard to show that this pattern holds generally, at least when n is finite. (Set theory has the interesting feature of allowing us to rigorously study infinities in mathematics, and this ever expanding cone of sets continues to grow through an infinite hierarchy of further infinities. But, uh, don’t worry about that for now.)
Back to visual design: The cork is printed with a depiction of V4, which is the largest stage of the von Neumann hierarchy that is interesting to visually inspect in its entirety. V4 has 16 members, but because of the way the pattern of growth involves exponents, V5 will have 2^16 = 65536 elements. You could draw V5 in a few weeks and check your drawing for errors in a few months, but it wouldn’t make a very interesting visual design element. (Maybe it would be neat wallpaper?) Incidentally, V5 is the last stage of the hierarchy that can be depicted concretely at all; following the pattern, V5 will have 2^65536 elements. You can’t draw this, even with lines one atom wide, because there are only about 10^80 atoms in the entire observable universe.