THE HOLES IN THE MACHINE
AN INTERVIEW WITH
LUCIANA PARISI
Beneath the machine’s surface is the source code: fluid chains of algorithms that determine its function
like DNA dictates the form of biological organisms, obscured from our view but increasingly shaping
our own environment as our experience of the world becomes more and more mediated through technology. Luciana Parisi is Senior Lecturer in Interactive Media at the Centre for Cultural Studies,
Goldsmiths, University of London. Her work is cross-disciplinary and involves complex and challenging speculation about planes of being that are inaccessible to the human, drawing on philosophy,
information theory and the sciences. In Abstract Sex: Philosophy, Biotechnology and the Mutations of Desire (Continuum, 2004) she engaged with the ontological and epistemological transformations that are entailed by the development of biotechnologies in cyber-capitalism and the incredible
complexity of bacterial modes of transmission and reproduction. In her forthcoming work, Contagious
Architecture: Computation, Aesthetics and the Control of Space (MIT Press, 2013) she turns
her attention to the abstract materialism of algorithmic objects: data entities that constitute an opaque
and ambiguous new ontological category, distinct from the biological and indifferent to the human. We
discussed this new direction in her research, along with neoliberal aesthetics, cyberpunk, the impact of
the technological on the human and the incomputable paradoxes at the heart of capitalism.
interview by NICHOLAS GLEDHILL
images of digital architecture by RE(MIX) S.A.M.S
NICHOLAS GLEDHILL: The first thing I want to
ask you about is in relation to the talk you gave at the
Signal:Noise conference earlier this year,1 where you
were talking about algorithms being not just executors
of programs but prehensive2 agencies that can evaluate
data and use feedback systems to act by themselves
within their own spatio-temporality. This conception
of algorithms is central to the work you’re doing at the
moment. Can you elaborate on it here?
LUCIANA PARISI: What I meant is that when I
started looking at what algorithms are, especially in
terms of the use of algorithms in design and in digital
architecture, what struck me is that algorithms were
not simply ‘stuff ’. There’s been a whole discussion
recently about how algorithms are the new stuff – the
new material that you can use to build with – and that
126: Nyx, a noctournal: ISSUE 7: MACHINES: SPRING/SUMMER 2012
instead of having bricks you now have data to construct
surfaces and buildings and urban spaces and so on. I
thought that this question was interesting, because
it wasn’t really addressed specifically in terms of the
ontological status of the algorithms. Instead I thought
that algorithms were not simply just stuff – just matter
that you can put together – but things, and as things
they’ve got an ontological status. So that’s one reason
why I started looking into algorithms as objects, objects
in terms of things. But then what does it mean to say
that there are algorithmic objects? Obviously there’s
a whole kind of return towards understanding of
objects [in metaphysics at the moment] as opposed to
surfaces, and this was interesting to me because within
algorithmic design and algorithmic architecture what
I saw was instead the way in which algorithms were
being used to generate surfaces, by being generative.
‘We are beginning to see this aesthetic everywhere now, for instance in architecture, the topological
model that is used is all about relations and continuity of change: the uniformity of change.’
So, let’s say you have sets of instructions, cellular
automata, and you produce a program by defining
what the instructions will do and then let them grow.
So there’s this idea of generative algorithms where one
algorithm generates another and another, very much
in terms of evolution.
as discrete objects, things in their own right, which
make decisions, and opposing it to the generative
model that you mentioned earlier which describes a
‘continuous flow’ of cause and effect, a kind of passive
evolution of algorithms constantly emerging one from
the other?
NG: So then, if in your view this concept of generative
algorithms in digital design and architecture is
inadequate, would this be because it doesn’t take into
account what you call the prehensive nature of algorithms,
the idea that they have an ability to make decisions
independently rather than simply being programmed
and then evolving according to set patterns? And could
this imply that for you they have a kind of agency, or
even a ‘consciousness’?
LP: Yes, in that model they’re constantly emerging
because one step is the basis for the next step, and so
on. So, it’s like you have ‘children’, it’s a very filiative
model, so you have ‘parent’ algorithms and then you
have their descendants. In this way the algorithm just
searches, is given a set of instructions that are left in a
space to proliferate, so what you get is these kinds of
images of these changing, morphologically changing,
forms that produce a certain kind of aesthetic and the
aesthetic that is produced is very much an aesthetic of
curves, of supple surfaces. We are beginning to see this
aesthetic everywhere now, for instance in architecture,
the topological model that is used is all about relations
and continuity of change: the uniformity of change.
LP: For me this is not really a question of volition
or consciousness but a matter of breaking up the
continuous flow or the deterministic patterns of cause
and effect. By breaking the chain of being, algorithms
are sorts of automatic prehensions that are constantly
making choices, evaluating, eliminating data. It is true
to say that these automatic prehensions can select
data that is not visible to us or even experienceable; in
short the point I want to make is that algorithms are
not instructions to be performed but are information
objects. In other words they are things, and not only
are they things, their ontological status also admits
that what looks to us like an automatism or simply
the culmination of formal reasoning instead exposes
a mode of thought that we cannot comprehend. For
algorithms the function of reason is not to verify
a theory or to construct a theory out of facts. It is
to calculate, process and quantify what cannot be
compressed into smaller units or programs. It is to deal
with the now constant production of ever-escalating
worlds of data. It is about considering data as objects,
things, real things, but also – and perhaps contrary to
object-oriented metaphysics – it is to admit that these
are things that think; i.e. that automatically elaborate
data, select and discern and ultimately take decisions.
NG: So you’re taking this alternative view of algorithms
128: Nyx, a noctournal: ISSUE 7: MACHINES: SPRING/SUMMER 2012
NG: Right, so this kind of thinking is reflected in an
aesthetic, and you talk a lot in your new work about
a kind of topology, an architectural aesthetic, which
for you represents a neoliberal aesthetics. In what
way is neoliberalism, or capitalism in general, linked
to this? In what way is the aesthetic that you critique
‘neoliberal’?
LP: One level is to do with emergentism, with a critique
of emergentism in so far as this assumes that from simple
things complexity emerges – complex form emerges, in
the same way as from DNA basic instructions you get
an organism. It is the same model that is at work in
this use of generative algorithms in design. So what’s
the critique of emergentism? The critique is to do with
the fact that, rather than thinking about a historical
process by which things happen, in emergentism
there appears a kind of spontaneity. And of course
you know the critique of spontaneity within political
debate has been quite strong because it’s a mixture
between vitalism and chaos; i.e. that things emerge no
matter what you do, no matter what is programmed
‘Algorithms are their own entity, and what they deal with is the incomputable data – something that
becomes so big that no system can incorporate it, not even capitalism. In a way I want to divorce
capitalism from technology.’
or planned, or strategically predetermined, things will
just emerge. So, there’s this idea that capitalism has
adopted this form of an emerging and self-producing
organism as it were, and that the aesthetic reflects this
image of capitalism, this tendency of capitalism.
aesthetic is something to be actively confronted and
opposed, or is it something that’s just inevitable? Or
perhaps we should view it as being an example of the
pharmakon that Stiegler talks about: poison and remedy
at the same time?
NG: So an image of capitalism in the sense that Deleuze
and Guattari talk about, with its tendency to constantly
deterritorialise and reterritorialise everything it comes
into contact with?
LP: Well I think that it’s more poisoning rather than
curing us, and if there is an ethic it’s definitely not
a neutral ethic, in fact it’s an ethic which neutralises
humans. What I’m trying to describe is the agency, the
agency of algorithms, and their capacity to produce
a spatio-temporality, or ontological existence that we
do not comprehend. So in a way this is to say not that
we are just acted upon – that we are slaves in the face
of neoliberal capitalism which will always already
reproduce itself – instead what I’m saying is that
capitalism itself has got its kind of internal paradoxes,
or internal tension, so whilst at one end the investment
in constant capital has become such that it has
produced, you know, automatic machines that think, or
algorithms that are for me thinking agents, on the other
hand you also have the fact that algorithms cannot be
controlled, that the investment of information has
turned against capitalism because information cannot
be completely computed. So there is some kind of hole
at the heart of neoliberal capitalism.
LP: Exactly, yes, and because of the way algorithms
are used to generate structure and the way they
continuously evolve you have this conflating
understanding of capitalism and technology. But also
that’s just one level; another level is to do with the kind
of organisation of urban space and aesthetic of urban
space where this kind of computational design has
allowed for the production of hyperconnection, the
potentiality to connect by creating smooth surfaces,
supple forms, continuous relations of points. So whereas
previously, in terms of the kind of capture of form by
architecture or the capture of the form of power, it
could be some kind of Le Corbusier house where the
rooms are cells and everything is really geometrically
organised, now you have this kind of flatness where
the ceiling becomes the floor and then you have the
open space and transparency and everything is kind
of merging together, and it’s this merging that I see as
being problematic, all this merging of points into one
surface. The smoothing out of edges has produced this
kind of aesthetic of the ‘blob’, an aesthetic of the curve
or of suppleness, which is something that Deleuze
and Guattari were already anticipating in A Thousand
Plateaus when they talked about the supple and the
striated.
NG: And when you say that this is problematic, that
there’s a problem with this, in what way do you mean?
What I’m wondering is whether there’s a strong political
or ethical critique intended here of these forms and of
this approach to algorithms and how they are being
used or whether you are instead just explaining how
this is happening. Is it your view that this neoliberal
NG: because of what can’t be computed, algorithms
which are no longer under control?
LP: Yes, in a way, in a way this is a kind of semiaccelerationist idea, if you know what I mean…
NG: ‘Accelerationist’ in the sense of a theory that
capitalism carries within it the seeds of its own
destruction; that if we just leave it to run its course it
will inevitably destroy itself ?
LP: Actually there are various strands of
accelerationism. There’s one that would just argue,
from Deleuze and Lyotard, that one just has to go
with the capacity of capitalism to deterritorialise the
human; and then there is another one that instead
tries to oppose – doesn’t just go along tactically with this
129: Nyx, a noctournal: ISSUE 7: MACHINES: SPRING/SUMMER 2012
kind of deterritorialisation but wants also to think of a
strategy. So, what is this strategic deterritorialisation?
What could it be? The question for me is to always
try to open the human to the existence of other
ontological objects, things or creatures. In my previous
work [Abstract Sex] it was bacteria, in this work it’s
algorithms, but what is important is firstly not to think
of neoliberal capitalism as one whole incorporating,
reproducing entity. And that’s why I argue against
neoliberal generative architecture, or generative
algorithms, because that kind of aesthetic and the kind
of overlapping of the technique onto capitalism just
says to us, conveys the idea to us, of this kind of megaorganism that’s always-already reproducing whereas I
think one has to oppose that. So that’s where I think
one could develop a semi-strategy that is not only
believing in the power of capitalism to destroy itself.
This does not mean that one has to fall straight back
into politics of resistance. What instead is important to
me is not the horizon of liberation, but the presence of
complexity all the way down.
NG: So would it be right to say maybe that the problem
with this kind of ‘aesthetic of the blob’ is that it’s
conveying a sense of inescapable homogeneity, a kind
of ‘fixing’ of the aesthetic and of the algorithms into
one amorphous form, a form that mirrors capitalism’s
view of itself and serves to hide the real complexity
and conflict under the surface, and that this fixing
process is negative because these are things that need
to be constantly re-approached?
LP: Yes, re-approached and broken down; in one
way to re-approach is for me to argue that algorithms
are their own entity, and what they deal with is the
incomputable data – something that becomes so big
that no system can incorporate it, not even capitalism.
In a way I want to divorce capitalism from technology.
There needs to be a constant re-thinking of this now
flattened relationship. My critique of these kinds of
generative algorithms in this responsive architecture,
interactive architecture, media design and so on is
exactly this: that they want to flatten the realm of
theory or thought with the realm of practice and doing,
because what you need to think is that this model that
wants to collapse technology and capitalism through
using the generative algorithm, what it does is to say
‘okay, we need the environment to reproduce, to act,
to react, to respond to us in order for us to produce
things’, in the same way as capitalism says ‘I need’ but
does it in a subtle way. Think of capital coming through
the windows of an interface, or an application, what
it wants is for you to act, to respond, to be included,
to participate, to interact, but for me that’s just a
façade, that’s the façade in which capitalism appears as
benevolent, the benevolent face of capitalism…
NG: Because you don’t really have any options?
LP: No, because all the options have already been
preset for you, it’s a probabilistic system where there
is no way out, or at least there seems to be no way
out because what you don’t see is the source code. In
fact the source code is behind everything, and it’s the
weak point of capitalism. It’s what capitalism cannot
incorporate because once you have the source code
you can reprogram the whole thing. So really, what is
important for me ethically, or strategically, is that the
theory behind things is important, and it can’t just be
flattened with bodies or embodiment, one needs to still
articulate the importance of thought and theory within
this tendency of capitalism to try and make it all about
what is felt: touch, the haptic, smoothness, bodily,
visceral, affective… I think one has to go somewhere
else. I’m thinking particularly in terms of responsive
or interactive architecture or media; it’s basically your
body, your capacity to touch, to feel, to be involved
that becomes the input for the program to add new
things to itself. So in a way we can flatten capitalism, or
neoliberal capitalism, with a generative or parametric
aesthetic, just like Schumacher’s parametric aesthetic
which he argues is this new style, a new avant-garde
style, but that people have argued against, saying that
it’s the style of neoliberalism. There’s an article by
Owen Hatherley in Mute Magazine on this.3 For me
it’s a flattening – this parametricism and algorithmic
architecture in neoliberal capitalism – but I think one
has to find the hole, the gaps within computational
capitalism, and these gaps in computational capitalism
are the incomputable. The more data are produced the
more they cannot be compressed; there is this entropic
tendency of information to be increasingly augmenting
in size and volume so it cannot be compressed in
any system, so in a way it’s a chaos in the heart of
computation, and in the heart of the neoliberal model.
NG: Right, and I think that can bring us back to
what we touched on earlier, which was your view that
131: Nyx, a noctournal: ISSUE 7: MACHINES: SPRING/SUMMER 2012
‘The cyberpunk vision is always interesting because of the fact that it’s “punk”, the fact that it’s dark,
it’s crude, it’s dystopian, the fact that it doesn’t give you any hope, this is interesting because it is not
appealing to the sole ontology of the human.’
algorithms are entities we can’t control or predict, that
are able to make decisions of their own. Doesn’t this
risk a kind of anthropomorphism?
LP: No, because what I am really trying to say is that
they are not biological creatures, they are information
creatures, which is another thing. They cannot be
anthropomorphised in the sense that they are nothing
to do with life or death and they are totally indifferent
to biopolitics as well. It’s just not biological strata I’m
talking about. In my previous work, I talked about
the biological strata and found holes in the biological
strata. Now I’m talking about the information strata,
and it cannot be collapsed onto the biological strata,
it’s something else. And the fact that they grow or
mutate or so on is precisely to do with software, like
when Gregory Chaitin is talking about metabiology.
Metabiology is not biology simulated by the computer
or by programs, it is programming itself having its own
processes and its own dynamics, because in a way what
is important, why it’s important to retain evolutionary
theories, is because they are quite rational. Evolutionary
theories have always challenged that idea of the divine
creation, they’re really very materialistic, a kind of
materialism. I want to maintain them because of that
materialism but I’m thinking about the information
strata. In terms of the idea of algorithms having an
intention or volition, I don’t use these terms. I use
a form of Whitehead’s notion of prehension. And
Whitehead used this notion of prehension to describe
not perception or intention. Prehension is more about
the grasping and evaluating, selecting and breaking up
data. That’s all it is, and for me it is this capacity of
algorithms being able to process data, to select data,
which is the capacity of algorithms to make ‘decisions’
that defines their mode of prehension. It’s like a mode
of thought but it’s an automatic mode of thought, so
in a way I’m kind of revising the notion of automatism
versus vitalism; you know there’s this debate, this
diatribe in evolutionary theory between Darwin and
Bergson for instance.
NG: So, then it’s a problem of determinism basically,
of whether the algorithms are automatic in the sense
132: Nyx, a noctournal: ISSUE 7: MACHINES: SPRING/SUMMER 2012
that they could only ever do one thing, or whether
they’re ‘free’, for want of a better word. Do you think
that these algorithmic objects might exist somehow in
between these two extremes? That they could explode
that binary and represent something ontologically
different?
LP: The binary between determinism and
indeterminism? Yes maybe, perhaps we can say that
they are semi-determinate; semi-empirical or maybe
quasi-empirical, and the reason why I’d say that is
because I look into information theory and information
theory is an extension of studies of the mathematical
and of the formal systematisation and axiomatisation
of the infinite. And what happens in computation is
that the infinite is not completely outside computation
but is just about calculable as infinite, so there is some
kind of computational entity, which is called Omega,
which is at once discrete – so you can define it, you
can understand the first few numbers of it – but at the
same time it is also infinite. So in a way for me this is
very important because this sense of the incomputable,
this kind of limit of computation, is not just some
kind of spontaneous ‘vital force’, it’s also somehow
determinate, or determinable to a certain extent, or as
Francois Laruelle would say, it is determinable in the
last instance, and thus it’s determinable as infinite. So
such a thing, a semi-determinate thing, can exist.
NG: Finally, I want to ask you about your view on
the debate about technology and the human, of the
impact of technology on the human. Obviously views
on this vary quite dramatically from dystopian visions
of a world in which humans have been supplanted by
machines – Terminator, The Matrix and so on, in popular
culture – to a belief in technology as an emancipating
force, allowing us access to new ways of being, exciting
new possibilities. Where would your thinking fit in to
this?
LP: Well there’s one philosophy that would argue that
humans and technology have always been together,
through a kind of structural coupling, a parasitical
relation between one and the other, like the idea of
extended cognition or other philosophical ideas of the
pharmakon or the parasite and so on, where technology
is almost a kind of necessary condition for change.
For me this is a kind of ‘autopoietic’ model – the
organism cannot live without the environment and the
environment becomes constitutive of the organism –
whereas I think one needs to get out of this idea and
accept that there are many other things of which the
human is not aware. Nonetheless, these other things
are always acting, and are constitutive of the real, but
whether we experience or are able to grasp some of
these kinds of ‘machines’ doing their own things or not
will not save us; even if we understand it or whether or
not we oppose it or try to control it. Because obviously
there’s been this tendency today of re-thinking nature,
re-thinking matter in terms of its artificiality: everything
is an object, everything is an actor, everything is on
the same plane. I don’t think that things are on the
same plane. One needs to get out of this mentality of
optimisation, I don’t think that increasing the process
of technological advance and acceleration can actually
derail the human or mutate the human, these kinds
of postmodern, cyborg visions are, well, I think we all
know now at this point that that’s not the way it’s going
to happen…
and that’s why I’m saying that the function of reason
must be revisited in its speculative form. If you think
about how the more science looks into matter, nature,
information and whatever, the more what comes out
is this existence of these semi-theoretical and semimaterial postulates that we can be theorised to a point,
that we can suppose work in such a way but that we
don’t fully know. The work that scientific epistemology
does is to reveal all these anomalies that cannot be
completely systematised within itself but remain like
an open question, an oblique gate into the unknown.
It’s interesting. What this also means is the irreversible
power of computation to produce data that cannot be
comprehended or contained within one overall system
of power or metaphysical system. If algorithms are
making decisions for us it is not to overcome us but
simply to be themselves. Algorithms are indifferent to
us.
NG: It is all a bit 90s now isn’t it?
1. Held at the Showroom in London in January 2012, Signal:Noise
II was produced in collaboration with Mute and Queen Mary
School of Business and Management. Parisi’s talk was entitled
‘The Speculative Reason of Algorithmic Objects’.
2. Following the use of the term by English mathematician/
philosopher A. N. Whitehead (1861-1947) to be ‘prehensive’
implies a capacity to grasp or apprehend information; to evaluate
data and respond to it in a way that may be to a greater or lesser
extent automatic.
3. See Owen Hatherley, ‘Zaha Hadid Architects and the Neoliberal
Avant-garde’, Mute 3 #1, Spring/Summer 2011.
LP: Yes! That’s not the way it’s going to happen,
otherwise we would have seen it; you know, the
transformation, the kind of cyberpunk vision. The
cyberpunk vision is always interesting because of the
fact that it’s ‘punk’, the fact that it’s dark, it’s crude, it’s
dystopian, the fact that it doesn’t give you any hope,
this is interesting because it is not appealing to the sole
ontology of the human. But on the other hand it’s
also true to say that technology is not really something
that’s going to be substituting the human. There have
always been things that are parallel to the human
but that we just don’t know. One important thing is
to acknowledge the obliqueness and the ambiguity.
Instead of saying that we can coexist with these things
and we’re all going to be fine, or on the other hand that
algorithms will supplant us, substitute us as another
entity; those for me are just scenarios that create a hype
and actually hide the fact of the real opaqueness, the
obliqueness of the existence of parallel universes that
we don’t know. Nonetheless, I must still say that I still
think the rationalist project of science is important,
Nicholas Gledhill is a postgraduate student at the Centre for
Cultural Studies, Goldsmiths, University of London and current
editor of Nyx. He is interested in human freedom and psychic
individuation.
NOTES
133: Nyx, a noctournal: ISSUE 7: MACHINES: SPRING/SUMMER 2012