2 Nov 2008

Summary of Alistair Welchman’s “Machinic Thinking”

by Corry Shores
[Search Blog Here. Index-tags are found on the bottom of the left column.]

[Central Entry Directory]


Alistair Welchman: “Machinic Thinking” in Deleuze and Philosophy: The Difference Engineer. Ed. Keith Ansell Pearson. London: Routledge, 1997.

Machinism is not Mechanism[1]. Minds are not mechanisms operating according to pre-determined relations, but are machines made of – and by – matter, expressed through intensive forces and their extensive effects. Engineering for Deleuze is not below the dignity of philosophy: Engineering is philosophy, because thinking is machining.

Efforts to engineer machinery have long been the aim of military armament. But Deleuze claims that war machines are not made by “royal science” but by nomad ingenuity: “less a matter of the materialization of scientific theorems than of patchwork and botching and bricolage and still somehow getting something right in the end, or if not, then not surviving to tell the tale” (213bc). Machines are makeshift jimmy-riggings spontaneously designed in the midst of raging battles, advanced by mad-minded experimentations and chance-effective gaffs and bozo mistakes.

Welchman claims that the more primitive notion of machine “reaches a certain apogee with Kant’s philosophy” (213):

“For Kant a machine is a system with at best only motricity[2] and not ‘formative force or drive’, and is therefore a system that acts as the ‘mere tool of external moving forces’ (Kant 1786: ‘Dynamics; General Observation’; Ak. 4: 532).” (221c)

In Kant’s definition, machines do not act by themselves but rather transmit activity given from outside it; in other words, they are not automatons, because they lack “motor force or design force” (221d).

In addition, machines have been stigmatized as dehumanizing and destructive, on account of their relations with rationalization, calculation, and Heideggerian techne (214a).

This dark vision of machines is seen in philosophy’s devalorization[3] of matter, which is defined negatively as what lacks in the non-material: “matter is motionless, lifeless, incapable of knowing itself, etc.” Locke claims that, “Matter by its own strength cannot produce in itself so much as Motion” (Locke 1690: IV. x.10); and Kant: “We cannot even think of living matter as possible (The very concept of it involves a contradiction since the essential character of matter is its lifelessness, inertia).” (Kant 1790: §73; AK. 5: 394). (pp.214b-c).

However, matter is also considered as “the pure (although unformed) act of Chaos” (214). Because matter in this sense is considered a threat, it is tamed with its mechanistic definition as inert (214-215). In this way, there is a “transcendent evacuation” of the term matter which empties it of content or effect. This subordinates engineering to epistemological, and the machine to an external force, and further, to a transcendent telos for which the machine is merely an instrument, and matter is made impotent.

But yet, matter and machinery seem to have something of an autonomy, as revealed by the fact that the engineering labs, and not the theoretical physics of academia, have produced the most striking and world-altering discoveries (215d).

Welchman claims there are two problems with Kant’s approach:

The first is that Kant fails to acknowledge the autonomy in the transcendental apperception. Deleuze says in Difference and Repetition that the syntheses culminate in recognition, “which is expressed in the from of the unspecified object as a correlate of the ‘I think’ to which all the faculties are related.” (DF 135). In order to synthesize any object, there must be a transcendental structure that presupposes the unified object in advance, this is his ‘object = x;’ and this yet-determined transcendental apperception is connected with a unified self, an ‘I think’ who is always the subject synthesizing intuitions. But for Deleuze, this transcendental apperception is not truly transcendental, because it is based on the operation of one same mind that is empirically given. For Deleuze’s transcendentalism of the self, there is not a unified I common to its diverse thoughts, but a developing self who is continually obtaining new incompatible features, that is, ever becoming something new.

The second problem is that the objects of Kant’s critique, World, Soul, and God, are not truly criticized, because they remain unaffected by the critique.

These two mistakes of Kant have the consequence that Engineering solutions to his machinistic problems become impossible given his subordination of engineering to science and his devalorization of matter. (216d)

Kant breaks from philosophical empiricism – which holds that nature is given – by claiming that nature is engineered, which implies that there is a unity of nature exterior to it, serving to define the transcendental. For Kant, desire is the “artisanal production” of nature (217a).

Here an engineering problem is posed: the way that nature is produced is understood by means of the application of pure a priori geometry and mathematics to nature. This results in technical machinism: “mathematically calculable science is presupposed; matter is thought transitively as the mere recipient of science; and engineering (Kant’s primary term is schematization) is thought just as the application or one-to-one mapping of science to its material object domain” (217b).

Nonetheless, Kant was unable to show exactly how science is capable of application, and was left always with an unaccountable “rebellious matter” which continually resists conceptual specification.

For Deleuze and Guattari, Kant fails to deal with the engineering problem of nature’s production by using a critique of machinistic thinking; what is needed is a “critique that demands that thinking itself become machinic” (218a).

This critique begins with a genealogy of the concept of matter. For long, matter has been thought inert, and “once matter is patient, and any capacities it appears to have must be referred elsewhere, then machines have become technical (aggregates with external sources of design or motricity) and engineering has become sheer application (of external sources of energy/design to material aggregates)” (218b).

By denying matter’s inherent activity, philosophy has refused to see positivity in matter and thereby repudiates immanent materiality. And because having already denied matter an inherent positive activity, when trying to do so, philosophy ends up endowing it with activities that were previously separated from matter, like consciousness, thinking, intelligence and so on. In other words, matter is not thought to have its own inherent positive activity (219).

This sort of critique is one that refuses to presuppose the constitution of the products when accounting for production; in other words, according to this perspective, what is produced cannot also be what did the producing, and thus material machines cannot have their own inherent force of production produce their very selves.

According to this theory, the functioning to a system is confused with what “impossibly overhangs systems and controls them from the outside”: in other words, the way a machine functions is not to be considered as being outside its functioning, like a hand turning a crank, for both the hand and the crank are part of the machine movements, as were the chemical energy of the body converted to mechanical, which came from the ecosystem, and so on; in other words, there is no closed system, and only when incorrectly presuming one do we confuse a machine functioning with something outside it.

But the problem is that matter cannot be homogenized, because matter is the explication of intensive differences. So there are strata: underneath extensive matter are a heterogeneity of irreducible intensive differences (219-220).

But even as intensities become extensive, these extensities themselves, on account of intensity, flee their strata or territory; they do not become inert matter permanently, but intensive forces come to shape them. This re-shaping of extensities by means of intensities is engineering or machining. It is for this reason that the Body without Organs is already machinic, because it is the product of the machinal intentive deterritorializing of extensities, while also becoming-machinic, because intensities as pure differences may cancel in extensity, but they are also implicated in themselves, which means they are never become exhausted. There is always another intensity within an intensity (220).

Thus there is no transcendental engine designer. What supports the system is the system itself. There are no “sky-hooks” as Dennett calls them, but just cranes which function as sky hooks. Cranes are machines sharing the same ground as the structures they build. The crane-hook and its construction are part of the same machine system rooted in matter and not in some transcendental support source (221).

It has been a mistake by Darwinians, Vitalists, and Schopenhauer to think there is an intentionality in the world: that genes mean to accomplish some ulterior aim, or that there is a will in the world. These are likewise sky-hooks. What drives the world is not ulterior to it, but materially immanent to it (221-22b).

Dennett takes this perspective further (than Dawkins), and proposes that even objects of human production, even the machines we make, are not human machines, but are products of the machines of which we are a part. We did not design our automobiles, Mother Nature did, by designing the human gene.

Guattari criticizes structural linguistics for being too abstract in that it allows all of language to be inter-translated on one strata of formal abstraction, but not abstract enough, in that it cannot represent non-linguistic elements on the same level as language (223c)

Deleuze criticizes the abstraction of representation as being too abstract, lacking “hands” that touch ground (223d). Deleuze and Guattari theorize that there are different strata of expression, with one being the “alloplastic stratum,” in which “not only is expression independent of content, but form of expression becomes independent of substance of expression” (224). On this alloplastic level, forms of expression become more autonomous, and abstract machines (for Dennett) become non-computable, as seen even in Dennett’s analysis of biological processes. Natural section is an “algorithmic process.” But it is a “heuristic” algorithm, and hence is non-computable. Also, natural selection involves randomness, which defies computational logic; and, this algorithm is not designed for anything in particular.

Machines deterritorialize when they become more abstract by spilling over into new implementations. So when human thought – which is the deterritorialization of carbon genetic replication into neural pattern replication – invents computational replication, or artificial life, this is again another spilling over of machine engineering. Hence contemporary AI efforts are Deleuzean engineering: “deterritorialized intelligence and life implemented in new media.” For Welchman, this is a “new alliance between Deleuze's machinic thinking and Anglo-American analytic engineering philosophy” (225d).

Thus Deleuze's philosophy should not be reduced to a tool for the humanities to produce critical texts. For Deleuze and Guatarri, machines are not literary tropes, but really are machines (226c).





[1] OED Mechanism, def.4: Philos. The opinion or doctrine that all natural (esp. biological or mental) phenomena can be explained with reference to mechanical or chemical processes.

[2] OED: motricity: Motor function; the faculty or power of movement by the body or a body part.

[3] OED: devalorize: trans. To lower the value of; to devalue. Hence devalori{sm}zation.

1 comment:

  1. thanks,this is great. can you say more about engineering? also would it be ok to say that what matters is also something positive by nature?

    ReplyDelete