Rendering(1): Recursive ray-tracing

Toward the beginning of Animal Capital, Shukin offers what she calls a ‘double entendre’ of ‘rendering:’

Rendering signifies both the mimetic act of making a copy, that is, reproducing or interpreting and object in linguistic, painterly, musical, filmic, or other media (the technologies of 3-D digital animation are, for instance, called ‘renderers’) and the industrial boiling down and recycling of animal  remains (p.20).

I have to disagree with both sides of Shukin’s double entendre – which I’ll call, for convenience’s sake, rendering(1) and rendering(2) - and indeed with the very notion that it can be so easily called double: as she herself recognizes, rendering “comprises much more than the logics of representation and recycling that I have singled out” (p.24), and hence I have to read rendering as a multiple entendre.

But I take issue even with her double formulation of the problem, even in spite of its recognition as an incomplete definition of the term.

In the first place, it is difficult to map a logic of mimesis onto the sorts of computational rendering practices that she herself associates with rendering(1). When a 3D animation studio renders an image, what are they rendering? If we imagine it as the process of converting a wireframe model into a virtually fleshed-out digital actor, then we can readily see a certain logic of mimetic reproduction at work. But when we recognize that this wireframe is itself a rendered visualization, and that in fact what a ‘render farm’ renders visible is a mass of code and equations, it becomes more difficult to refer to this unambiguously as a form of mimesis. Although such rendering is no doubt often  inhabited by a mimetic logic – as in the case of motion-capture – it seems more like a case of transduction than representation or ‘interpretation:’ the displacement of information from one medium to a very different one.  Rendering, in the first sense, reconstructs a model of a scene from structural information encoded in machine-readable language, and then produces an image intelligible to humans by simulating the interaction of virtual light with that scene (by any of a variety of mathematical techniques: ie, raytracing or scanline rendering). Shukin’s association of this sense of rendering with mimesis - and the rhetorical gymnastics that ensue – reads like a reductio ad absurdum of a certain representational, ‘literary’ logic. To me at least, it seems far more sensible to conceive of rendering(1) as a basically non-representational, non-mimetic procedure that is occasionally inhabited by an older logic of mimesis than vice versa.

But really, I’m more interested in rendering(2), since it was the occasion for me to actually do something for this week’s posting. It seems curious, albeit predictable, that Shukin places such exclusive emphasis on industrial rendering, for the most part relegating mention of more artisanal practices to her critical engagement with the attempts on the part of industrial renderers to rhetorically justify their practices with reference to these ancient ones (eg. Shukin p. 59). This is a justified target for critique, to be sure; but again I want to problematize her ‘definition’ a bit. I read rendering(2) as signifying, in the first place, the melting and breaking down of animal fat for consumption – a sense attested from 1792 – with the practice of industrial rendering, the object of Shukin’s critique, being a secondary and derived (albeit nevertheless important) sense of the term.

With all of the fetishization of so-called slow food that’s been circulating lately, there’s been a growing interest in artisanal rendering practices familiar and accessible to the home cook – following one of Natasha’s interlocutors from Modeling Proteins, the sort of rendering that one’s grandma might have done with chicken fat. (That rendering, incidentally, would produce schmaltz, a tasty product commonly used in place of lard by those keeping kosher or halal.) Thus you have bloggers who’ve rendered lard, books about the joys of animal fat and rendering at home, and a general backlash against the whole anti-fat kick of the late 20th century.

In keeping with this trend, and looking back to the videos we watched on “slicing, dicing, peeling, and palpating,” I’ve documented in photos the production of my favourite dish involving rendered hog fat: carnitas! Spanish for ‘little meats,’ this is essentially just a fatty, well-marbled and well-exercised cut of pork – I use bone-in leg or picnic shoulders – braised for a long time in its own rendered fat.  The videos we watched for slicing-and-dicing-week really reminded me of this procedure in some ways (but not others, which I’ll address a bit below), with all the predictably creepy associations that implies. But I still enjoy making and eating it. What follows is thus rendering, reflection, and recipe – although what it shows is in some respects the exact opposite of what you see in your average recipe book. It may not be for the vegetarian or the squeamish, but it’s no more unsettling than your average dissection video.

Here I’ve just begun slicing the skin off a 3lb half picnic shoulder, a cut taken from just above the animal’s front leg. The bone you see in cross section is the leg bone. Carnitas, like Southern barbecue, use tough muscles like these that are full of connective tissue, because they’re suited to a long cooking time at a low temperature: this allows all of the fat to render out and much of the connective tissue to melt down, making the meat much more flavourful (and, of course, less healthy) than a lean cut.  When slicing off the skin, the goal is to neatly separate the skin and subcutaneous fat from the underlying muscle – this means you have to pull the skin back as you slice. As you can see, I’ve left some meat on the skin. While pulling back the skin, my hand brushed some of the few remaining coarse hairs and I shuddered a bit, throwing off my cut. This is a fairly common occurrence when I take apart cuts of pork like this.

Here I’ve finished removing the skin, and begun disassembling the meat. Again, you have to pull and separate the segments of the muscle as you cut, following the lines of connective tissue and intramuscular fat. I’m very picky about such things, so I tend to remove all of the fascia and blood vessels that I can find. The white fat will all render out as it braises, though, so I leave that.

I’ve finished separating out the musculature into pieces of a good size for braising – for the most part following the lines of the connective tissue, but also slicing large pieces into more manageable portions. The meat is seasoned with a mixture of cumin, salt, pepper, and smoked paprika, then doused with lime juice. The salt and the acid in the lime juice start in on the meat right away, denaturing proteins and breaking down collagen. The bowl on the right in the second picture above contains scraps – the sorts of bits that would certainly be used by an industrial renderer, but which just gross me out. (Again: picky.) The bone, and its accompanying bits of muscle and collagen, I keep in the freezer for stock-making.

The rendering process begins: I slice the skin and its accompanying fat into manageable pieces, and place them in a hot cast-iron frying pan. They fry for a while, then I add water and cover. I simmer the skin alone for about an hour or two, adding water when necessary – you can tell by the pitch of the noises coming from the pan. Once it shifts from soft bubbling to high-pitched sizzling, I add more water. Doing this for a couple of hours gets me enough hot lard to sauté the meat, but in order to actually render out all of the fat in the skin, one would have to simmer for at least 24 hours.

Now, finally, I cook the meat: first it is browned in the pork fat, and then I add a braising liquid made up of water, more cumin, paprika, and salt, and a handful of dried chiles. Beer is a good addition if you have it on hand, but on the day in question I wanted to reserve said beer for drinking with the carnitas. After the braising liquid is added, I placed the skin back on top, so that the fat would continue rendering out and keep the meat moist. Then you simply place the pan in the oven and braise at around 300 degrees for as long as you can stand it… In this case, I think I let it go about four hours before I had to dig in.

Here’s the end result, after pulling-apart with fork and fingers:

The traditional way of serving carnitas is with fresh corn tortillas, lime, cilantro, and crumbled queso blanco. This time, I felt like burritos. Nachos are good too – generally speaking, spicy pork braised in its own fat is a pretty versatile ingredient.

Make of this rendering what you will – personally, I’d recommend that you make dinner of it.

More seriously though, I wanted to document in a concrete way how industrial rendering is not the singular sense of ‘animal’ rendering. I was also struck, after putting this rendering together, how even though this gives you all of the information you’d require to prepare the dish, it faces you with precisely the messy, muscular physicality of culinary labour that’s often totally absent from traditional recipe books. (By way of contrast, take a look at David Lebovitz’s beautifully-photographed blog entry on carnitas.)

At the same time, the hog that this shoulder came from was no doubt produced by the same agricultural-industrial complex at which Shukin takes aim in her text, and those parts of it which were still less valuable than my $1/pound picnic shoulder likely found their way to an industrial rendering plant. So my intention was by no means to undermine those reasonable and necessary elements of her particular critique… but only to raise some questions about the so-called double entendre of ‘rendering’ and some more of its multiple senses.