Tuesday, May 10, 2005

Jukka-Pekka Kervinen Interviewed by Mark Young

M.Y. With your father a composer you have a history of music; the depth & complexity of your computer programming skills indicate a long period of interest & work in that area; but you seem to have sprung full-blown as a poet, like Botticelli's Venus from the half-shell. I'm guessing here, but a reference to 'mesostics' in a footnote to some poems that appeared in
Shampoo #10
tempts me to open by asking if the influence of John Cage had anything to do with your moving in to the writing field.

J-P.K. One correction: my father was an artist (painter and comic arts), my brother (and now my oldest son) are composers. But it is right that my childhood/history was mostly musical. Cage and his textual work was one reason why I started to move to the writing field. Cage's mesostics as well Mac Low's acrostic and diastic writing had a great influence and were a starting point for me. Anyway my first attempts had more to do with musical composition. At the time I was planning to do a composition with prose notation and started to examine more closely the work of American contemporary/experimental writers. I found much more than I had expected, people like Jim Leftwich, John M. Bennett, Peter Ganick, Michael Basinski, Sheila E. Murphy and others. And that was all I needed! Instead of prose composition I started to make experiments with writing (without computer). They were published in 'Moria' in 2002 under the title 'meditations', then gradually I moved to work with a computer using the same programs I already had written for working with compositions. Those Shampoo #10 texts were my first attempts with computer and they use Cageian mesostics as their underlying structure.

Had you ever thought about writing poetry previously? Was music still giving you satisfaction? The reason I ask is because your
'Deep Listening'
pieces, which also come from around 2001, seem to me to be a transition between notional & verbal. One of them, Breath, which I quote below, seems to echo the 'manual' poems that appeared in Moria.

Breathe fully,
inhale - let abdomen expand,
exhale- abdomen contracts.
In every 2, 3, 5 or 8 exhale
make a sound in any way,
mostly ppp-p, soft,
occasionally ff, short,
more silence than sound in duration of exhale.
Close your eyes, be relaxed, listen to others.
There is no synchronization,
every performer has their individual breathing rhythm.

Duration of piece is free.

In public performance, use dim, green light.

J-P.K. Yes, I had, first time at age 12 or so when I found e.e. cummings works in a local library. At that time he was an exact counterpart of Anton von Webern for me, I saw so many similarities between them and tried to write poems in spirit of cummings. Next time ten years after, again we must jump ten years and then in the year 2001, when I started to write 'seriously'. It wasn't that music didn't give me satisfaction, rather it was another view from my starting point. I have always been interested in structures, any structures, and I don't see much difference whether the content or surface is aural, visual or textual. I just do same things with whichever material I choose to work with, any of three I mentioned (which often also overlap very much !). Poetry has all three. By that I mean in textual works the visual appearance is very important, much more important than in music, and of course (even when reading in mind) its aural structures are central, poetry is anyway meant to be read aloud, I think. Those Deep Listening pieces are really from the same 'source', they were made at exactly the same time as first 'moria' poems and the content in both is same, Meditation which I have been interested in (and practised) for a few years now. And 'Breath'-piece is in prose notation, it was one of those works which eventually led me to write poetry/textual works.

M.Y. Was there an overlap between the programs you used for your computer-generated music & what you used for the poetry/textual works? Were there text-manipulation programs out there that were available for re-engineering? Or did you know what you wanted, & worked out a way to get there?

J-P.K. Yes there was an overlap. The first experiments I made were with the same programs I used in computer-generated music/scores, then I wrote a few programs which used acrostics as their base structure. After that I started to write programs and utilities to use only for textual manipulations. Sometimes I use same algorithms as in composing programs, but usually they are quite specific, meant only some for specific operation or group of operations.

I found some programs on the Internet, but I wasn't impressed by them. They didn't do the things I wanted, usually they tried to generate poems by imitating some style with bad results. I wrote and still write all programs myself, it is the best way for me. I am not a professional programmer, not at all, but have been interested in programming languages and their theory and studied quite a lot of programming languages. In textual manipulations I have used several languages, each of them having their own 'identity' which affects the way I manipulate texts/images.

M.Y. I'll return to the programming aspects a little later, but let me just go back a bit. Your first computer-generated pieces used poetry – Gertrude Stein, Alan Sondheim – as a source & they were presented in what I'll call a poetic format. Did you feel it necessary to proceed this way, that in order to 'validate' what you were doing you needed to work within poetic parameters? & receiving what I'm guessing was a strong positive response did you then have the confidence to range much further in both source & output?

J-P.K. I started by using Stein and Sondheim as you mentioned and much of it had to do with 'validation'. I was in unknown field, but still I didn't have any program how to proceed. I just fooled around and picked anything I found interesting. As I practically started from 'scratch', I printed most of the materials I found from Internet (I didn't have any other source !). After four or five months our bookshelves was full of printed papers and when I cleaned them there was 4.5 meter bunch of printed material ! Although the sources I used produced more 'poetic formats' it
still wasn't what I was looking for. I sent a few of first experiments to some e-zines (Poethia, moria, shampoo) just hoping to get some response to my works, really didn't know how to validate it in a 'larger context'. And you're right, they all accepted my works (which wasn't my purpose at all !) and their response was very positive. I was happy, and confused ! But of course I was more confident and I decided to move further, to use those procedures which better describe my nonlinear thinking. And one of the first things was to give up using existing works by other poets.

M.Y. It's almost a paradox, giving up poetry & poetic formats to produce poetry. Did you have a vision of what you wanted, where you wanted to go? & how did you know when you got there, when you reached that point where you could say "this is truly my own"?

J-P.K. So far I have known exactly where I have wanted to go, not necessarily how, but I have had a vision of direction and goal. But of course there happened unexpected things. I have found very interesting things before I have reached my goal, things which have changed my course and interest to somewhere else. Also, there is no system or formulae to say when I have done it, it is still a very intuitional thing, it is feeling, and usually very powerful. Many plans I have made have been disasters (I think), but still I have found something which has been very useful somewhere else. Although I use the computer to make choices, I still need (and want) to decide most things, including when texts/images are ready to go.

M.Y. Okay, let's move on to the how. & the why. & I'm only talking about your text works here. I'll get to the visual images later.

Putting aside your internal processes for the moment, there are three parts to your work. There is a computer program which, by definition, has to be precise; there is the output which is determined by chance though bound by the parameters & the probablity distributions you have included in the program; there is the source material which can be randomly or deliberately selected, these days mainly randomly I think. How do you choose which of the three to address first? How much backtracking do you do? (What I mean by that is how often do you rewrite the program or change the source material if you don’t like the output?) How do you go about it?

J-P.K. Yes these are the components. First is an idea: many times a kind of model, textual "class" where to start. It is usually very precise, I know exactly what it has in it. Next phase is to make a computer program to simulate it, to produce that text type. 99% of backtracking is with program, usually I have several versions of same program only slightly modified. I continue until it produces exactly what I want. I never edit source material. After using Stein's poetry, Sondheim's texts, Wittgenstein etc. I made my own source vocabulary and I used it over two years. Now I have made a new source, it is larger than the first and have used it a year or so. Occasionally I may use Marx's texts or other poet's but it is quite rare. The programs themselves have usually several parameters to keep them (and output) flexible and to have more control for the output and behaviour of software. Most of the programs are based on chance operations or maybe a better expression would be composer John Myhill's "controlled indeterminacy". But there are also others based to feedback, recursion, information theory and nonlinear systems. But the starting point is always textual, these are only means to simulate, process and produce the output.

M.Y. When you say you "know exactly what it has in it" what exactly do you mean? Are you talking about the spatial arrangement, the internal form, of the end piece, or its external appearance? Are you talking about whether words are entire or punctuated by space(s)? How you will use punctuation marks, either from the source or as substitutes for words? Those sort of things?

& how do you know when you've achieved what you want? If you're working with "controlled indeterminacy" then how can you be sure that another iteration won’t give you more? & "controlled indeterminacy" itself. My interpretation of the term would be that at a certain point there could be several pre-determined next states, but which one is actually chosen depends on chance. I seem to remember reading of a piece by John Cage where what happened next depended on what instruction card was randomly picked out from a number of cards. Is that what you mean by the term?

J-P.K. By "to know exactly what it has in it" I mean that I know both the spatial/visual appearance as well internal form/ relations between elements. I must know, it is not possible to write computer program for generating text/images without knowing what you try to achieve by that program. Many times I write preliminary version(s) by hand, then I start programming to simulate text I have written. For me it is a very interesting way to work, trying to find rules, patterns, constraints, schemes from the existing work. When you make explicit programs, even using stochastic techniques, it means that you still have to write an explicit computer program, every detail must be planned and programmed, the program will not and shouldn't generate anything apart from what you yourself have put there.

Choosing the iteration is maybe the hardest phase. Usually I do it just by intuition. If I have made 'a model' which I simulate, it is quite easy to see when the work is ready. But which iteration is better or more precise is a problem I haven't found a solution for. Except one. If I use stochastic methods for text production, it is possible to make final judgement stochastically. For instance in my first book ,
each piece/text/poem is sent to a listserv automatically ie. I don't choose, I decide when program is ready and then I send next (or nth) iteration straight to mailbox/listserv. Without reading it beforehand. For me this 'method' is most consistent with the way I generate pieces. But that's not the only method I use, there are other posssibilities which are based to same kind of stochastic 'criteria'.

M.Y. & the "controlled indeterminacy"? How do you decide what random variables to use in a particular program? How / where / when do you use them?

J-P.K. It is always an idea, what I'm trying to generate or achieve with any particular program.
Different probability distributions fit different situations. For instance normal or Gaussian distribution can be easily controlled by two parameters, one for mean and second for variance. If the event space is constant and I want to generate elements from some area of it, I just set the mean to a point in the middle of that area and variance to its ‘borders’. Also, event space can variate and change as well as probabilities. Distributions are a very flexible way to control generating elements whether they are letters, words, lines or musical notes. But I use also other methods, some of them are not random at all like nonlinear chaotic systems or some recursive functions. They are deterministic systems, and even though they are not random their output may be hard or impossible to predict. And many of them are surprisingly simple!

M.Y. Can you clarify for me how you use distributions. Let's use normal distribution, since it' s one that most people are familiar with, the bell-shaped curve with the mean, median & mode coincident, the standard deviation is the square root of the variation & 68.26% of the items are within the range of the mean ± 1 standard deviation with lower concentrations between the first & second & the second & third standard deviations. I have no problem with this in regard to height or weight or temperature, but how does a word become a random variable? How do you determine the mean? Is it the input or the output or both that you choose – or consider – to be normally distributed?

J-P.K. Of course! As a method for generating/manipulating text with any distribution the most crucial thing is how to define event space, ie. the material from which you're generating. Word as such is not sufficient (usually). As I previously stated it can be anything: words, lines, forms, musical notes etc. If I use words it can be a source text, dictionary, programming code source etc. But I have used it with individual strings too. Let's take an example: my e-book
(PI 2004) consists of three versions or iterations of same series of programs. Each version ends same way: the text goes toward letter 'n' gradually. This was generated (or first version of it was) by normal distribution by setting mean to letter 'n' (or to index of it in normal alphabet) and variance to larger number (like the number of normal alphabet/2), then setting binomial probabilities to choose whether printing letter from source string or letter from normally distributed letters. After that I gradually change variance of normal distribution toward one and binomial probability to toward normal distributed letters. It results text consisting of only letter 'n', but slowly and 'smoothly', iteration after iteration.

Explaining these things may sound complicated, however I think they are still basically quite simple. My mind just works this way, has always worked, I am constantly seeking new approaches to generate text/images and my thoughts organize things like this.

M.Y. Okay, let's move away from ' lines' & on to ' line'. Your father - & I've got it right this time because you corrected me earlier – was a painter, but I remember you remarking to me in an email somewhere that you were too impatient waiting for paint to dry to become one yourself. & yet, obviously, painting & drawing have always been important to you. I'm thinking here of the covers of the xPress(ed) books that you publish, a series in an early eratio that was both line & word. If we look at your
MailXart blog
which is where I think you express yourself most purely as an artist, I see influences of Mondrian, Schwitters, Miro. I know of your love of Lichtenstein from our use of one of his pieces on the cover of
The Oracular Sonnets
. What finally prompted you to branch off from the purely textual to text & colour?

J-P.K. I have been interested in art from my childhood, many artists but also art theories and styles.

Although I approach any poetry first aurally and then visually, they are equally important for me. Any poem has visual aspects and tendencies. My textual works were already quite visual, intentionally, many poems were actually based on visual ideas. And moving to text and colour was a logical consequence: if I think textual works as a two-dimensional plane it is restricted (because of nature of letters and their positions) by row/column-thinking ie. discrete positions. I just eliminated that moving to continuous planes which are not restricted by the fact of columnar thinking. Textual forms in two-dimensional plane consist of say 80 columns (ie. letters) and ca. 60 rows (ie. lines).Visual forms are not restricted (theoretically) at all, I can put any element to
position 440,216 (x/y), and if I want to move it slightly I am able to use position 440.0416,215.998844. A minor change, but it is essential, that makes a plane infinite. Columnar thinking is of course finite (but it is very interesting as such, it still offers endless possibilities).

M.Y. Can you tell us something of the genesis, the growth of these visual pieces. The precursor, for me, is a
black & white post
that appeared on your Nonlinear Poetry in May last year, that seemed like a magnified segment from a textual piece, large letters that overlapped, about which I commented at the time that "the moon falls from the sky & rolls towards me", but which struck me as having been constructed through a Word Art program even if its content had been arrived at through a stochastic program of your own. What did you have in your mind, in your mind' s eye, when you created this? & was this the first, or had you been working on them for a time before this?

J-P.K. I started to work with visual pieces in 2002, published a few pieces in eratio and Generator, wrote more programs until I was 'ready' in last year. Pieces starting from May 2004 are all done with combination of some "conventional" programming language which produces Postscript (so there is no Word Art program...) which is again converted to some image format. With these works I started to explore a "new" two-dimensional area (it was new for me...), the shapes of letters, limits of cohesion, nonlinear maps, among others. There were so many things I had (have) in my mind, each image constructed a type (class) of images to explore, to experiment and elaborate more. I worked with Postscript until now, but recently I have started to use an image editor(s) for another kind of works, before that I haven't used any image editor, again a new way to work, new methods and things to study. Extremely interesting!

M.Y. I don't know the Generator pieces, but the
eratio ones
seem to me to be text with asemic drawings added. What strikes me about the pieces that started appearing in May – apart from the advanced conceptual aspect, again fully-developed like Botticelli's Venus – was how integrated, how cohesive, they were / are. I have always approached them as works of art because of the painterly qualities they display; their structure, design, balance, tension, etc. Most times the text is secondary for me because of the way it phases – to use a word from
Gregory Vincent St Thomasino's insightful piece
on his eratio auxillary blog – in & out, is form rather than traditional signage, & yet the words & letters seem to imply that the pieces are there to be read – whatever interpretation we place on "read" - & this is reinforced by those pieces where we recognise the text. Your development, directions, new paths, have been evident since then, branching off into new blogs, separating textual pieces from visual pieces; but there is little existing evidence of how you arrived at this, to borrow a project term, May milestone (though knowing how you work, I'm guessing that privately there is). & your answer above is not all that forthcoming. Could you please tell us more about the genesis of, the philosophy behind, these pieces?

J-P.K. I have been interested in asemic writing for a long time, both Jim Leftwich and Tim Gaze have defined most of its theoretical foundation I know, fascinated by the idea of writing "having no semantic content", something between writing and not-writing, writing-like gestures with just emotional/visual content. So I started doodling and found things when I wrote very fast, trying to do it without much thinking. Works started in May are computer simulations of those doodlings,
working same way as I simulate my own texts with computer. Some of them are "original" works in ink on paper, it was playing with conscious mind (computer) and unconscious (paper and ink). Very interesting, challenging in its own way, meditative working. I also see them more visual than textual but still they remind me of writing as action, without semantic content. Now it seems that these asemic writings have melted to my all visual works and I haven't updated my 'asemic' blog much.

M.Y. This may or may not be a question, or several questions, but I just want to make sure that we're not talking at cross-purposes. Do I understand from your previous response that the text & colour pieces evolved / devolved from your asemic work, that the line(s) or the space(s) between them were replaced by items of text or areas of colour? Or have either of us mis-interpreted the other's question / answer? Are you talking only about your asemic pieces, or about all the initial pieces?

Actually asemic works are quite independent although my working methods are same in text/color pieces and in asemic works. But I was talking about the asemic pieces. As with any other domain of life I am interested in having multiple approaches to any particular issue and the asemic works are one issue. Those pieces represent to me the pure emotional aspect of writing, the unconscious, the uncontrollable, the mystic ritual.

M.Y. After the last few words of that response I feel guilty about doing this, but I want to pursue the text/color pieces. What gave you the idea for them, who / what influenced you, how much experimentation went on before you felt you were ready to show them? I keep coming back to this because of their completeness, how fully-developed they were right from the start. They were brilliant, but unexpected. I can find no hint of them in the body of your work before this time. Or maybe I'm missing something, have overlooked a clue.

J-P.K. Before the first 'eratio'-pieces I experimented a great deal, made hundreds of pieces before I was able to do the pieces I started posting to nonlinear poetry. I don't think that you've overlooked a clue, for me the transition from discrete two-dimensional plane (textual works) to continuous plane (visual works) was a leap into the unknown, yet logical as I always felt that I want to move freely in that plane. I don't know if emphasizing visual layout in my textual pieces can be seen as a clue, anyway with both planes (of course they are physically same) I am still working with same structures, as I have always done.

M.Y. We've barely touched on the other aspects of your creative life – mail art, editing, publishing, collaborations, photography, your reasonably recent involvement in the Finnish literary scene - but we're running out of time & space. This interface – until recently your only one - with the outside world is obviously of critical importance to you. So, as your last response, could you process your thoughts stoically &/or stochastically & give me a summary, an overview of these activities & what you get from them, what you feel you give through them. & two final questions. How do you find the time to do everything you do at the pace you do it? &, what next?

J-P.K. All these activities are very important, crucial, in my life, I try to devote my whole life for them. I am working the same themes with different approaches, each of them having a different aspects and methods of working, still issues are basically same, chance, (controlled) indetermination, interpretation, the relationship between writer and reader, and structures and systems. Finding time is always a problem, by devotion I mean exactly that, I have focused on what I consider essential (writing/publishing) and have pruned everything else. My life is very simple, I work in the middle of my family, 12-16 hours in a day, practically just writing and editing. Still I feel I don't have enough time, unfortunately, there are so much interesting things waiting for exploring. The future is a great, beautiful mystery. I don't make plans, I just walk forward and pick things I encounter on that trip. And I am grateful of everything I have found so far, and everything I will find in future.