Petrified Media: opening talk

This is the text of the talk that I gave at the opening of my solo exhibition Petrified Media, at the Earth Art Gallery in Bristol. Some of it may find it’s way into an article that I’ve been asked to write for Cultural Politics, but for now here are my thoughts about the residency I began in November 2019, that was supposed to last 6 months and is now wrapped up more than 2 years later.

At its core this project is about the relationship between rocks and digital image technologies, or as I now think of them – earth media and technical media. Everything we make, every cultural object, is produced (or grown) from the resources beneath our feet. By changing the physical state of rocks we have made hammers, cars, iPads and Rheometers. However, over the two year duration of this residency I have come to question the anthropocentric perspective of such statements that confers decisive agency exclusively upon human actions. Geologist Peter Haff articulates an alternative – that technological agency is not entirely subordinate to human agency. Haff conceives of a technosphere, with equal planetary impact as the atmosphere, lithosphere, hydrosphere or biosphere. However, Haff writes, “unlike the other paradigms, the technosphere has not yet evolved the ability to recycle its own waste streams”.

Coming into the School of Earth Sciences two years ago, these were the sorts of questions I began the project with (some of which will likely seem absurd):

– What is the difference between a rock and a camera?

– If a camera can be made from a rock then is the rock not already capable of photography?

– And what happens when a camera turns back into a rock? (as is presumably already happening at some unknown depth in the Earth’s crust). Does it return to a previous lithic state or will it sediment new minerals to be unearthed by future geologists?

It was questions such as these that led to the statement that opens the video I made during the residency: “To photograph a rock is to point polished pebbles at a rock”. Which is really just to say that the image sensor in our cameras is a sliver of near-pure silicon, extracted from a predominantly silicate crust, refined, and wired to a series of also-silicon computer chips to process light into image data. But, haven’t rocks always been processing light into image data? The photosensitivity of silver and silicon are not inventions after all, but discoveries. Benjamin Bratton writes that with the appearance of the first photograph of a black hole the aperture of the camera has become the size of the planet  – the image being triangulated from simultaneous telescopic exposures at multiple sites. But hasn’t the earth – absorbing solar radiation and developing a photosynthesised bio-image across its land masses – always been a camera?

Speaking with Heidy for the first time I was struck by the centrality and continuity of silicon in the experiments being conducted on the disequilibrium project. Silicate rocks are collected from volcanic sites, they are imaged onto silicon image sensors at Diamond Light Source and these images are analysed and the data input into a computer model in Manchester – an algorithmic magma chamber, synthesised in silicon, in which the behaviour of liquid silicates can be modelled. By the time I return from my first trip to visit the various scientific partners in the consortium, this reflexive arc of silicates within their experiments has captivated me. To summarise it in a single phrase: Minerals are melted into machines to analyse minerals while they melt.


One of the central problems addressed by the Disequilibrium project seemed to me be to develop a methodology to photograph a rock while it is melting.  I turn these words over and over in my head:

       To photograph a rock while it is melting.

       To photograph a melting rock.

       To melt a photographic rock.

       To melt a rock while it is photographing.

and resolve that my response should be to melt the camera.


But my camera is mostly plastic, and no one takes photos with a camera anymore, so perhaps inevitably, I turn my attention to my recently broken iPhone 5, the camera of which also has a scalar relationship with the samples used by the volcanologists. Heating a contemporary technical artefact to a temperature higher than the melting points of its components reveals its underlying materiality. Melting breaks down the temporary arrangement of these materials into a functional whole, forcing apart alloys and rearranging elements according to their physical properties rather than their electronic schematic. Anode and cathode seep together. Neat metallic squares of micro-components, once located to control electrical currents flowing between them – now flow together: cobalt, copper, titanium and zirconium recrystalise with one another in a matrix of molten aluminium that previously enclosed them in commodity form. A molten phone, baked in the subterranean heat of a nearby magma chamber, is one potential future. To melt a camera then – is to manufacture a technofossil.

Jan Zalasiewicz has written extensively about contemporary technofossils, the likely traces of our Anthropocene era that will be evident in the stratigraphic record of the planet to future palaeontologists. Silicon and quartz are so inert and resistant that, in his opinion, they are most likely to defy the chemical weathering of deep time, perhaps even more so than the industrially hardened types of steel found in many consumer technologies. It is possible then that some of the computer chips now embedded in almost every electrical device might survive the extremes of temperature and pressure, and that the microelectronic paths etched into them will retain or imprint their form in the surrounding bedrock. Graptolite fossils have survived to the present as the hollow spaces left by their skeletons became pyritised (pyrite, for the non-geologists among us is fool’s gold). In commenting on the likely candidates for pyritization among our current urban detritus, Zalasiewicz identifies the interiors of any of the myriads of tiny metal and electronic gadgets that we now produce in their millions … for these in themselves contain iron, one of the ingredients of pyrite”. Even this brief discussion of the futurity of media objects, it is clear that the samples I produced during the residency will bear no resemblance to the real effects of deep time on today’s e-waste. Isolated from the hydration of the subterranean environment and dramatically accelerated in comparison to centuries or millenia of gradual baking and compression, the furnace is a blunt instrument whose results are in no way comparable to the speculative future they seek to materialise. Yet this is also how Science operates, by removing samples from their context and simulating the forces upon them in a controlled environment. When I began the residency in November 2019 I was given a tour of the building that navigated its facilities from the surface to the core, according to the depth of the processes synthesises by their apparatus. But science has not yet designed an apparatus for the re-crystallisation of media into minerals, hence – as Haff states – there is no means to metabolise the inherent waste of the technosphere.

In another essay Zalasiewicz uses the example of this object, known as the Antikythera Mechanism, which was found 120 years ago in the Aegean Sea inside a 2000 year old shipwreck. 72 years later its purpose was understood by analysing it with X-ray tomography, revealing it to be an analogue computer engineered to predict astronomical positions and eclipses, the toothed cogs which enabled its identification also appear to us now as a 2000 year pre-echo of Charles Babbage’s Difference Engine. It is somehow inconceivable that our current technologies of computation could be forgotten, reinvented and then rediscovered in as little as 2000 years, but it has already happened in the short time of human history. And when computation is reinvented it use the same techniques?

To tomograph an object is to image its interior in slices or sections. Volcanoes and computers can both be understood by tomography. At Diamond the volcanologists rotate a high-pressure furnace in the path of a high-energy X-ray, observing pyroxene crystallisation in real time before reconstructing a 3D model of crystal growth in a chip of pure silicon. I remain incredulous that it is possible to study an object of the magnitude of a volcano in a space smaller than the camera in my phone. Tomographing that camera allows me to scroll back and forth through the object, to excavate its not-yet-fossilised form. Which of these shapes will erode? Which cavities might be filled with pyrite? And which might survive the chemical weathering of deep time?

Tomography has become an epistemic tool applied across disciplines and scales. A few years ago while researching the role of photography in the Fukushima clean-up I came across an experimental technique called muon tomography being trialled to image the interior of the melted-down reactors. Muons are produced by the collision of cosmic rays with particles in the upper atmosphere, and are capable of penetrating deep into the earth’s crust. At Fukushima, scientists built an instrument capable of measuring the frequency and trajectory of muon strikes. The theory being that as uranium is dense enough to scatter muons, the nuclear fuel should theoretically produce a shadow in the resulting photograph. So, from one perspective, the barrage of cosmic rays perpetually striking the planet and all of the structures built upon it, can be conceived of as a kind of imaging. We have always been being tomographed, continually imaged from every direction by penetrating radiation.


For me, the tomographed image of the camera is interesting in and of itself, but for the scientists working on Disequilibrium the image is valuable only once its contents have been analysed. Observing the scientists work on this project I became fascinated by another epistemic technique which seems to suddenly be everywhere: segmentation. Segmentation is the process of annotating or labelling an image, and is a vital precursor to all machine vision systems. Nolwenn, one of the post-docs employed at Diamond on the project, was spending much of her time segmenting different crystals within the sample. But segmentation has become ubiquitous and is also performed as a kind-of piece-rate digital labour by a globalised workforce to whom image annotation is outsourced by services such as Amazon’s Mechanical Turk. Cameras and images can become operationalised within automated systems, but only with reference to dataset that has been ascribed labels by a human user. I can hardly imagine a more emphatic illustration of Haff’s conception of the technosphere as “a system that operates beyond our control and that imposes its own requirements on human behaviour” than this decentralised precariat obediently drawing boxes around every pedestrian in an image database to underwrite the eyesight of autonomous cars. In this inversion of anthropocentric perspective, a human life is only valuable once it has been segmented.

After a period of lockdown I return to the samples of molten iPhone cross sections in the basement. I process the resulting lumps of molten metal using the tools and techniques of the petrologist, grinding and polishing them for microscope imaging and electron microscope analysis. Media theorist Jussi Parikka speaks of A Geology of Media to draw our attention to the deep time planetary processes that produce the ores extracted in the service of our industries and sciences. The pounding of waves that has ground and sifted monocrystalline quartz for centuries before sand is scooped into furnaces and stretched into fibre optics, the gradual accretion of heavy minerals in the coastal sands of Western Australia and Senegal from which Zirconium is refined for the manufacture of nuclear fuel rod casings, or the coursing of thermal springs through volcanic pumice that precipitates a lithium-rich brine now pumped from beneath the Atacama desert for our phone batteries. But, in practice, the dark grey ingots of no-longer-smartphone are relatively unyielding to the experimental tools of geology. Perhaps to work with the material aftermath of technology requires more a Metallurgist of Media.


According to a project conducted at the University of Plymouth where an iPhone was ground to dust and X-ray diffraction performed on the results, a smartphone consists of 22 different metals:

33g of Iron                              

13g of Silicon                         

7g of Chromium                    

6g of Copper                         

2.7g of Nickel                         

2.5 g of Aluminium               

1.6g of Calcium                     

0.7g of Tin                              

900mg of Tungsten                

160mg of Neodymium          

90mg of Silver                        

70mg of Molybdenum           

70 mg of Cobalt                    

36mg of Gold                        

30mg of Praesodymium        

20mg of Tantalum                 

10mg of Niobium                  

7mg of Antinomy                   

5mg of Gadolinium               

2mg of Dysprosium               

2mg of Germanium               

2mg of Indium                       

To identify a rock using the traditional method of optical mineralogy it must be sliced and polished to a thickness of 30 microns or 0.03mm. At this thickness it becomes translucent and its index of refraction can be measured by the interference pattern of polarised light shone through it. To identify a rock then, it must first be turned into a lens. The rock is incorporated into the body of the camera; earth media becomes technical media. The screen of a phone is similarly polished, buffed smooth to the molecular level using the rare-earth element Cerium. However, the refined metals contained in my manufactured slices of iFossil cannot be identified by optical mineralogy, so, I use the electron microscope. I am told that looking for rare-earth elements in molten lumps of iPhone is literally like looking for a needle in a haystack. But I persist and eventually find that in addition to the 22 metals identified in Plymouth the fragments of white lattice found in this section of iPhone screen heated to 1000ºC are made of pure Zirconium, and that the contacts around this capacitor contain Bismuth as well as gold and silver.

Among the temporalities of geological processes such as sedimentation, metamorphosis, and compaction, the assembly of a mobile digital technology of the early twenty-first century occurs dizzyingly fast and with a startling planetary reach – sourcing metals from every continent. We might even go as far as to describe these mutually alien timescales as existing in a state of disequilibrium. Volcanic activity is often produced by the geologic process of subduction, in which one tectonic plate is pushed beneath another. Subduction zones are sites of geological recycling, folding portions of the crust back into the metabolism of the lithosphere. But there there is also a state of disequilbrium between this planetary capacity to regenerate material and the technosphere’s inability to metabolise its own waste.