Paul De Marinis „According to Scripture“ (2002)

In the present project I have drawn a number of illustrations and engravings of sound vibrations from old physics and acoustics texts, many of them predating the invention of the phonograph, and translated them into short sound files. None of these sounds have ever been heard since their brief transit from vibration to inscription in the physics laboratories during the years 1853 to 1890. These are in effect, the most ancient recorded sounds now available to our ears. (1)

The title derives from a play on the name of the noted 19th c. acoustician, E.W. Scripture, remembered most for his enlargement of the grooves of phonograph recordings via an elaborate pantographic apparatus.

Some of the recordings are of patently banal sound sources: sirens, tuning forks and the like, and were created to form a kind of baseline for the visual interpretation of sound phenomena. They served to calibrate, verify and unify a host of laboratory practices and experimental apparatus with mathematical and theoretical studies. Other traces were intended to be objects of study in incipient fields, such as phonology and linguistics, and offer unique glimpses into particular sound moments from what is supposed to be a silent era. While none of these is comparable in import to the microwave echoes of the big bang, they are sometimes astonishingly clear and accurate, however brief. Their very existence and their present decoding may offer another interesting footnote into the the history of the recording media; also, certain artifacts characteristic to them present some interesting questions and speculations in the ongoing debates about the transition from human writing to machine writing as it applies to media theory.

The philological efforts of the 18th century bore much fruit in the era that followed. Monumental efforts like Champollion's that succesfully decoded lost languages served as proof that secrets long hidden could indeed be deciphered. That the romantics saw in the traces of phenomena the writing of Nature is well known; that this writing might be read and interpreted was the project of numerous scientific endeavors of the nineteenth century: stratigaphy and paleontology, fossils and the origin of species, spectrography and the signatures of the elements. Among early efforts to fix the traces of sound, the figures of Ernst Chladni drew much attention early in the 19th century, including the accolades of Napoleon and the mathematical attentions of Sophie Germain. These two dimensional patterns were created by lycopodium powder dancing on a metal plate that was drawn into vibration with a violin bow. As it happens, these figures, though visually replete with meaningfulness, turned out to have more to reveal about the shapes and attachments of the metal plates than the sound itself. Still, the hunt was on for a species of automatic writing that would uniquely reveal the inner nature of sound vibrations and fix them for systematic visual scrutiny and measurement. The phonograph that was to provide verification of ts own data by reproduction was still a distant glimmer.

A host of machines and methods were devised to fix the traces of sound, among them optical devices. Part musical instrument, part camera, part chart recorder, they appear as the mechanical equivalent of the chimeras of ancient legend. The manometric flame, an invention of the prolific Charles Wheatstone, used a rotating mirror to display the variations in the height of a flame excited by sound waves as a time sequence. These images figure prominently in several of the early texts, but the precise method by which the visual observation was fixed and transfered to printing plates in the texts is uncertain. The phonautograph of Leon Scott in 1857 was a major breakthrough in the automatic translation of sound vibrations to script. A horn concentrated sound on a membrane coupled to a hog-bristle. The bristle rested lightly on a rotating drum covered with a piece of paper lightly coated with lampblack. As the drum turned it advanced minutely on a screw creating a long helical trace that wrote out the variations of air pressure as slight wiggles in the line. Fine wiggles were created by high frequencies, broader strokes by lower ones; larger deviations from the center line were the result of louder sounds, etc. This gibed well with the currently gestating theory of sounds. In fact, combination tones could be deciphered as regular and evolving patterns waves that could be visually analyzed to reveal their constituent components, offering proof that a single point could indeed convey multiple superimposed sounds. Scott foresaw in the ramifications of his machine a whole new science and postulated its use in linguistic studies and even as the basis for a new literature, all the time focusing on the handwriting-like quality of the phonautograph traces.

These traces presented incontivertible evidence of the nature of sound, the basis of a scripture of nature that would promise to be above doubt, interpretation and beyond the reaches of error and fraud. The simplicity of the phonautograph seemed to insure that its product was as real as could be. Yet as we now decode and listen to these ancient sounds, it is evident that errors crept in to these automatically dictated writings of divine nature. These were manifold: misidentification is one reason; since there was no way to replay the traces and verify the content, traces shown in these pages are sometimes incorrectly identified in the texts, for example the (ratio/interval) in plate (plate ) is actually a (ratio/interval);

Other misidentifications are more suspect. Within the re-translation to sound carried out in my project it is quickly apparent that the traces of (name) "Hello how do you do... Brown University" are completely spurious and contain none of the sounds purported by the author.

We can sometimes hear the artifacts of the recording machines themselves (experimental apparatuses)the trace in (plate#) betrays the glissando of a machine slowing down to a stop.

Another interesting sources of error is not readily apparent to the eye, but emerges as we process the images to convert graphic into sound files: In some cases the traces seem to go backward in time, however briefly, and confound the ability of my algorithm to decide which of two simultaneous points is actually the right one. The origin of this error can be traced, I suspect to the confusion of machine writing with handwriting, for the traces of the phonautograph had to be manually converted from scratches in lampblack into reproducible figures on copper plates.

The work of converting machine generated images to drawings and then to engraving was the work of cadres of trained engravers whose skill was to create reliable verisimilitudes of the figures presented to them by the scientists. Nonetheless, certain the hand of the engraver emerges from the page to re animate the machine-generated script. Many of these images were copied from one edition to another , thence to translation and popular knockoff. Certain conventions of hand drawn lines may be responsible for the creation of these impossible time or direction-reversals. See plate (plate#) and (plate#) details.

The disembodied hands that attend the instruments in engravings of demonstrations of the phonautograph and related sound-writing devices are another source of curiosity. These experimental hands, a hallmark of 18th century representation of scientific experiments, (see (name, paper)) persist in a number of these illustrations. We may wonder if the engravers, recognizing the similarity of their work and the apparatuses and processes they were being employed to represent, felt some connection with the phonautograph's needle. Sometimes amusingly gendered (see plate (plate#)) into engraver and engravee, we wonder if the hands depicted might be less those of the gentleman-experimenter but echoes of the hand of the engraver who created the plate, like Escher's drawing hand.

As a scientific genre this particular style of graphic traces of sound persisted well into the 20th century, and even in our own time adorn the pages of scientific and acoustic journals as well as the windows of our digital sound editing software. Some of the sounds in my project post-date the invention of the phonograph, but are deciphered and played here because of their contibution to the history of sound. In particular Harvey Fletcher's hallowed "Joe took father's shoe-box out" was a real delight to hear as its scratchy tones proceeded from the silent pages of the book to the scanner , through hand corrections in Photoshop, then Max/MSP to the raspy voice on my loudspeaker. Other traces, like the manometric flames posed a series of vexations. We know what they are supposed show about sound with respect to frequency and amplitude, but they also record artifacts of flame combustion dynamics and infrasonic variations. Already forsaken as potential revealers of vowel color the manometric flame was completely abandoned as a tool for illumination of sound quality as soon as superior photographic methods such as Dayton Millers phonodiek superseded it. Even more recent traces like the Russian "Aaah" that present an alluring pair of physical and acoustic replicas, offered challenges not so much because of their obscure coding systems, but because the conventions of halftone reproduction, to which our eye is favorably accustomed, make the definite line that the machine can follow quite fractal in nature. Many of these brief blips are sounds only a mother could hear.

I made a number of decisions and judgements regarding my own interpretive licence in presenting these traces as sound. In some cases I chose to use my hand and eye to interpolate apparently missing portions of the trace. In other cases, such as „Aaah“, I applied injudicious amounts of graphic processing to force a result that was finitely decodable. Throughout this process I was aware of my own interventions, sometimes practical, sometimes playful, as an ongoing part of laboratory and studio practices that have really never ended for these traces. In offering them here for audition and speculation my intention is to pass them on to further experimental and interpretive recycling as sound samples.

The efforts of the Romantic era , as Kittler has pointed out, were focused on reading and writing, not on the recreation of primal sound. The mechanization of writing, though, eventually suggested, even demanded that these scripts be played anew. That the earliest mute records treated here are now reduced to their electronic media equivalents, processed & played by computer, is evidence that the great project of automatic writing, reading, deciphering and hallucinating proceeds like the well oiled machines of yesteryear.

(1) is the sound of an existing Roman era bell, for example, dating from the third century a more ancient sound? For this to be the case we would have to think of the bell itself as an encoding of some „sound“; that sound, in turn, would have to include the splashing of molten brass, the beating by smiths‘ hammers etc. But the sound the bell produces in its current use is far from being a recording of these sounds. In my scheme no sound exists naked and no moment of sound can be recorded without ist attendant accompaniment of stray vibrations in this gedank example in the form of strike-tone (the material with which the bell was struck plus the force and neccesarily the quivers of muscles imparting that force, muscle-tone having ist own history(cite Sci Am article)) and the admixture of sounds from the environment. Therefore for the purposes of this projects I do not consider the hypothetical bell itself to be either the substance or the carrier of a more ancient sound.

Technical note: The traces were scanned on a flatbed scanner, extracted and isolated by a number of processes in Photoshop, then transformed into audio files via a custom patcher in Max/MSP. The sounds were then presented in the exhibition as .aiff files played back on a conventional CD player.

(c) Paul DeMarinis 2002