Digital nominalism. Notes on the ethics of information society in view

of the ontology of the digital

Tere VadenDepartment of mathematics, statistics and philosophy, University of Tampere, 33014, FinlandE-mail: tere.vaden@uta.fi

Abstract. The commodification of code demands two preconditions: a belief if the existence of code and asystem of ownership for the code. An examination of these preconditions is helpful for resisting the furtherwidening of digital divides. The ontological belief in the relatively independent existence of code is dependent onour understanding of what the ‘‘digital’’ is. Here it is claimed that the digital is not a natural kind, but a conceptthat is relative to our practices of interpretation. An interpretative system that sees code as something that canor should always be owned implies an increase of social control and threatens vital processes of knowledgecreation that are necessary for an open and egalitarian information society. The ontological belief in ‘‘digitalcode’’ thus provides the backdrop for an ethical view of the information society. Consequently, if we see digitalcode as an interpretative notion (in the nominalist way), the ethical questions appear in a different light.

Key words: digitalness, digital divide, disorganisations, free software, ICT, information society, intellectualproperty, knowledge creation, ontology

Introduction: The digital information society, code

and postmodern philosophy

The concept of the digital contains a plethora ofhopes and fears. A new way of ‘‘digital being’’ isthought to prompt both the need for new digitalontologies and epistemologies, as well as to resultin novel ethical dilemmas. These expectations havea real effect when the so-called information societiesare given shape by concepts like ‘‘intellectualproperty’’ and ‘‘digital rights management’’. If thecentral feature of the information societies is therole of digital information as a tool and as aproduct, then the definitions and interpretations ofterms like ‘‘information’’, ‘‘knowledge’’ and the‘‘digital’’ are of essential importance. Crucially,digital information allows for new configurations ofbasic social forms like property and knowledge.Because digital information can be reliably andcheaply copied, it has great potential for demo-cratisation and equality of information distribution.On the other hand, digital technology also makespossible the monitoring and controlling of infor-mation so that the commodification of knowledgeis possible in ways that threaten to create hugeasymmetries or so-called digital divides.

The ontological idea of the ‘‘being of the digital’’or of ‘‘digital being’’ is closely connected to the ideaof the immateriality and pervasiveness of informa-

tion. This idea of disembodied information, imma-terial knowledge and artificial intelligence (AI) is acommanding tenet in the tradition that sees a utopiain ‘‘digitalness’’. The utopia of AI as well as theutopia of digital being are based on the idea thatsome processes or entities in the world (digital,algorithmic, informational, intelligent, cognitiveprocesses; in a word ‘‘coded’’ processes or entities)are defined and given existence by their abstractforms in a way that makes the details of their physicalconstitution irrelevant.1 The coded processes aremultiply realisable in different physical make-ups.The utopia rests on separating two different onto-logical and epistemological areas. On the one handwe have the abstract and functional world of digitalcode (information, representations), on the otherhand the concrete and obstinate corporeal world ofphysical instantiations. These two worlds, the expe-rienced physical world and the idealised abstractworld, are thought to be isomorphic, and this iso-morphism is though to be the basis of our knowledgeof the abstract world.2

Along with the ontological and epistemologicalquestions, digital information processing raises

1 As a level of explanation this coded existence is called func-

tionalistic: one discusses the functional roles of abstract entities in

separation of their physical realisation; see, e.g. Block (1980).2 The idea of isomorphism goes back at least as far as Plato.

Ethics and Information Technology (2004) 6: 223–231 � Springer 2005DOI 10.1007/s10676-005-0350-7

ethical questions on the global scale, not the leastbecause of commodification of code (intellectualproperty rights, information patents, patents onDNA, copyrights on immaterial assets etc.; for areview of current legal and social developments, seeDrahos and Braithwaite 2002). The commodifica-tion of code and, consequently, of knowledgeembedded in code needs two preconditions. For thefirst, what is needed is a sustained belief in theexistence of ‘‘code’’ in which information contentcan be expressed, transferred and through which itcan be identified. This is the ontological side of theissue: a belief in code as a more or less indepen-dently existing entity. For the second, one has toset up a system of ownership for this code. This isthe socio-political, or, more generally, ethical sideof the issue. In this sense, the first condition doesnot only comprise of a scientific way of looking atthe world and explaining phenomena as if theywere in the last instance algorithmic or ‘‘coded’’. Italso provides the necessary background for a par-ticular view of the ethical implications of informa-tion society development. To put the matter in acrude simplification: if code did not exist, it couldnot be owned. Consequently, different views andinterpretations on exactly how and where code doesexist imply different views on the ethics of infor-mation societies. ‘‘Code’’ is a both scientific term (atheoretical term, if you wish) with an epistemologyand ontology of its own, but also a governmentaland juridical concept invested with economic andpolitical interests. It is becoming clear that the firstcondition is closely coupled to the second conditionin a world where economic production is dependenton intangibles. For example, ‘‘intellectual property’’as a concept is a way of controlling and organising– in the Taylorist sense – the new type of pro-duction.3

If it indeed is the case that the notion of‘‘intellectual property’’ functions not only as a basisfor a kind of information society but also produceseffects that widen the digital divides between own-ers and non-owners of information and knowledge,it is of essence to note that the processes and theconcepts behind them – ‘‘code’’, ‘‘digital’’, ‘‘prop-erty’’ – are historically and socially contingent. Thedevelopment of information societies is in partdependent on how these concepts are understood

and employed.4 One of the most debated andprofound themes in 20th century philosophy wasthe controversy over the nature of technology. Onone extreme, technology was seen as a neutral,harmless and helpful tool for humankind to use, onthe other extreme, it was seen as a pervasive powerout of human control and with pernicious goals ofits own. Marxist thinking, Popperian rationalism,critical theory, and Heideggerianism all have theirown distinct views on the matter. The questions ofthe non-neutrality of technology, in general, and ofthe interests included in the use of a particulartechnology become salient if we continue to takethe question further by asking, in a postmodernvein, what structures of power are embedded inthese interests, what productive effects (in theFoucaultian sense) they have, or, more bluntly,who or what groups of people benefit from a par-ticular use or form of technology. The interestingquestions about technology, including digitalinformation technology, can only be formed whenthis kind of ‘‘who’’-analysis (analysis of the struc-tures of power) is combined with the insight thattechnology is not one thing, not one identifiablewhole with a lasting essence or drive leading toparticular kinds of society. ‘‘The same’’ kind oftechnology (its forms and use) may now benefit thisgroup of people, the next day that group of peopleor form of life. So the questions have to be moreconcrete: what forms of power does a particularway of using technology support, need, presupposeor undermine? What are the forms of subjectivityor community included in a particular understand-ing and use of information technology? What kindof technology does a particular social regime sup-port? These questions are available only if bothsociety and technology are seen as non-essentialisticentities, with no immutable or everlasting features.

Through this set of questions technology loses itsneutrality in both of the senses: it cannot be seen as apure tool or as something that forces something onus. Naturally, technology is also a tool. However,understanding a particular technology as a tool has ineach instance specific consequences. Equally, tech-nology as a way of thinking is beyond our will aspersons. But these two extreme positions only makethe questions more pressing. These questions link theuses of technology into the manifold and contradic-tory social, cultural and political developments andpower structures that develop also for reasons rela-tively independent of technology. These power

3 Copyright legislation and the laws concerning other so-called

immaterial rights are being changed at a quick pace. This trend

is most marked in the US, but the EU is following close behind.

It has been claimed that ‘‘never before have so few owned so

much a part of intellectual property’’ (Lessig 2002, see also

Lessig 2001); these words do, intentionally or not, have a

familiar ring.

4 For an interesting historical account on how technology

influences our relations to documents in the general sense of the

word, see Levy (2001).

TEREERE VADENADEN224

struggles and technological interests are visible, forinstance, in the practices organised by the concepts of‘‘code’’, ‘‘digital’’ and ‘‘property’’.

Code and its existence

Two ways of resisting the commodification of code

The hypothesis is, then, that what was called ‘‘com-modification of code’’ above is a combination of twoinformation technological interests: (1) the belief in afunctioning layer of code (to be separated from thephysical instantiation) and (2) the set-up of a systemof ownership for the code. If this is the case, two waysof counteracting the development of an increasinglyasymmetric information society present themselves.One straightforward stance is to accept the existenceof code (to believe in it, and act accordingly), todefend it as a way of understanding the world, but atthe same time insist that the ownership of code leadsto blatant contradictions, as well as to social injusticeand economic imbalance. One of the ensuing con-tradictions is maybe most evident in the case of sci-entific knowledge (including theories represented informal code, such as mathematics, large parts ofnatural science, programming, etc.), which receives itsspecial status and credibility from the very fact that itis not owned: knowledge becomes scientific onlythrough the open and free critique of the scientificcommunity. To quote Jacques Derrida: ‘‘[...] in ascientific text [...] the value of the utterance is sepa-rated, or cuts itself off, from the name of the authorwithout essential risk, and, indeed, must be able to doso in order to lay claim to objectivity.’’ (2002, p. 47).As a speech act, a scientific text has to be distin-guished from the person or persons who ‘‘sign(s)’’ it,otherwise we are not dealing with a text that canassume the special characteristics and authority of ascientific text. This has been and still is largely theway in which scientific information and knowledgeare severed from a concept of private property that isdependent on the link between a person and anentity. The author, the one who ‘‘signs’’ science, is thescientific community. A particular way of speaking, aparticular type of speech act, namely that of scientifictexts, creates a community and a way of appropri-ating knowledge that is different from the case ofprivate property (as understood, e.g., in the Lockeansense).

A similar device for co-operating without theintrusion of private property has been developed inthe case of computer software. So-called free softwareis developed by a community of people who share-and-share-alike: the goal is to develop software that

the user is free to use, modify and redistribute pro-vided that the freedoms are transferred. In this sensethe ideal is close to the ideal of science. For thispurpose the movement needs a legal and social tool,one that uses the copyright claim set on a piece ofsoftware for community building rather than privateproperty building. This tool, developed by RichardM. Stallman, is often colloquially called ‘‘copyleft’’:the copyright statement in question gives the user theright to modify and redistribute (the modified versionof) the software provided that the right is alsotransferred (see Stallman 2002). This viral nature ofthe ‘‘copyleft’’ copyright protects the informationand knowledge amassed in the software frombecoming closed by ownership. The knowledge isappropriated inside the common control of thecommunity.

In both cases, science and free software, the con-dition of the practice, as well as its goal, is a com-munity of sharing based on a certain set of commonvalues and interests. Both can be seen as ways ofacting, as power structures, that are instrumental increating an information society that contradicts thetrends of commodification of code and knowledge.As such the practices of these communities alsodemonstrate that digital information processing (orany other technology) does not force us treat code asproperty.

The second way of resisting the development of anincreasingly asymmetric information society is todirect our attention to the first condition, to theprocess of seeing the most salient features of theworld as code and, especially, to the peculiar kind ofexistence that code has. This questioning can be setapart from the questions concerning ownership, butmy aim in the following is to connect the questions ofthe existence of code to problems of the social andcultural status of code. Thus the ontological prob-lems are intended to guide the view on the ethicalproblems of information societies.

Digital existence

How does ‘‘the digital’’ exist? The digital is not aprecisely defined technical, philosophical or scientificconcept. It is also important to emphasise that thedigital is not a natural kind: it is impossible to build a‘‘digitalness-detector’’ that would start tickingwhenever something digital is present. The devicescalled digital are so dissimilar in their structure,physical construction and functions that distinguish-ing them from other types of devices is very hard.Also, the distinction between the digital and non-digital functioning of a digital device is a tenuous one.For instance, considered as a physical entity, a digital

DIGITALIGITAL NOMINALISMOMINALISM 225

computer functions essentially according to the sameprinciples when it does what it is supposed to do andwhen it doesn’t (due, say, to a broken wire). Theword ‘‘digital’’ is essentially an engineering conceptthat has to do with the way in which electronicdevices are supposed to work. To be more precise,‘‘digitalness’’ is a demand on the reliability of thefunctioning of a device.

One way of approaching the particular type ofexistence that the digital has is to look at the oppositeof the digital, the analog. A quick and coarse defini-tion is to say that analog devices work on continuousscales and are sensitive to the whole scale of a relevantfactor, while digital devices contain discrete jumpsfrom one state to another. The Latin ‘‘digit’’, toes andfingers, includes this possibility for jumpy and bitwiseoperation. One can count by using ones fingers torepresent discrete numbers, jumping from 3 to 4without going through the continuous scale from 3 to4. Letters are digital in the same way. Where one letterends another starts. A definition of the exact point oftransition is not necessary as long as the letters can beseparated and recognised reliably enough.

The efficiency and alluring beauty of digitalness isa direct follow-up of this discreteness and step-by-step interpretation. At least in principle it is possibleto repeat a process of a digital device at will in exactlythe same way (for example, the transition from 3 to4). To be more precise, it is possible to repeat aprocess so that the features considered relevant arepreserved. In this way it is possible to copy, store andreproduce a digital representation. The word ‘‘worm’’can be printed arbitrarily many times by using dif-ferent typefaces and forms of writing but the copiesare in a sense (the sense of representing the abstractcontent ‘‘worm’’) perfect as long as the letters ‘‘w’’,‘‘o’’, ‘‘r’’, and ‘‘m’’ can be reliably recognised. It isprecisely this (functionalist) sense in which digitalrepresentations are independent of their physicalinstantiation, independent of the media. This is howthe supposed disembodiment of the digital codestarts, and the abstract immaterial existence begins.In general, digital systems are those systems theprocesses of which can be repeated at will withenough precision and reliability so that the processescan be recognised as being identical with each otherwith regard to some relevant characteristics.

All definitions of digitality will be contested, andcontain different biases, tensions and, consequently,are non-neutral. However, what concerns us here isthe ontology of the digital and the effects that ontol-ogy may have for the ethical questions around digitalinformation technology. John Haugeland (1998) hasoffered a useful definition of digitality. A digital deviceis defined by two processes: first, a read–write cycle

and, second, a system of signs in which a type-tokendistinction can be maintained. A digital device readsand writes tokens of sign types, where the type isdistinct from the physical tokens, e.g., this ‘a’ is atoken of the type ‘‘a’’. Like discussed above, thetokens and types must be distinct and mutuallyexclusive, a sign is either an ‘‘a’’ or a ‘‘k’’, not both.Digitality can be achieved when the read–write cycleof the signs is reliable enough, that is, when the writtentoken of a type is recognised (read) as an instance ofthe type. Like Haugeland points out, the decisivefactor is the reliability of the whole cycle: the writingprocess can be more lax if reading is precisely con-trolled and vice versa. So, according to Haugeland, adigital device is composed of three things: (a) a set oftypes of signs, (b) a set of processes for reading andwriting tokens of the types, and (c) a set of conditionsunder which the read–write cycle is reliable enough.

With technology like this, it is possible to produceperfect copies of sequences of sign types (not ofsequences of tokens). In the sphere of the abstracttypes, a digital copy is perfect, even though all thephysical copies (consisting of tokens) in the empiricalworld are non-identical. Because the level of typescan be separated from the level of tokens, the (rela-tive) matter- or media-independence follows.Accordingly, it follows that not everything in theworld is digital. A major part of the devices in theworld, including a large part of those that have beendesigned to work digitally, are not digital. Condition(c) is in practice often the one that fails to be satisfied:digital devices get broken because the read–writecycle is not reliable enough. A digital device stopsbeing useful for its intended purpose when throughwear and tear it deteriorates to a point where thetokens it presents get blurred so that they can not bereliably recognised anymore.

The distinction between types and tokens is crucialhere. If everything matters, even the fact whether an‘‘a’’ is an ‘a’ or an ‘a’, we are dealing with somethingnon-digital. For example, causal relationships in thephysical world are non-digital (and analog) in thissense: they are sensitive to the whole continuous scaleof initial conditions. The movement of a billiard ballcan be explained by relying on abstract laws ofmovement, but the movement taking place in theworld is sensitive to the initial conditions in a con-tinuous and non-digital way. There is no upper limitto this sensitivity, therefore physical (analog) pro-cesses cannot be perfectly copied.

Digital relativity

The issue can be discussed in more detail by consid-ering the question whether everything that is analog

TEREERE VADENADEN226

can be digitised. Because the digitalness of a device isalways relative to a function or a goal, that is, relativeto the reliability needed, the digital is ontologicallyspeaking always dependent on something else andtherefore insubstantial or epiphenomenal: we alwaysexperience a physical world of particular tokens. Thedigital is constructed, abstracted or interpreted out ofthe physical. For instance, when faced with a digitalcopy of the Mona Lisa, we direct our attention onlyto certain aspects of the light emitted (those thatresemble the light reflected from the painting) whiledisregarding others (including what the light tells ofthe physical substance emitting/reflecting the light).Some properties of the painting can be digitalised(e.g., properties of the light reflected), while others,like the material doing the reflecting or the uniquecausal chain from Leonardo’s brush to the wall of theLouvre, cannot. The causal chain or the causalproperties of the painting cannot be digitalised simplybecause in those cases we do not know how to sep-arate between the abstract layer of ‘‘code’’ and thephysical layer.

Consequently, everything is not amenable to digi-talisation. For instance, even if we would agree thatsome kind of level of code is at work inside somephenomenon, if that code does not support a dis-tinction between types and tokens, digitalisation ofthe code is impossible. To put the matter bluntly, ahug is not digitalisable, because we do not know whatwould be the abstract type of which a particular hugis a token. A hug is dependent on unique causalchains, including the physical presence of two personsenduring the same physical risks, threats and prom-ises.5 This is another way of saying that things thatare dependent on particular causal chains and par-ticular times and places can not be digitalised,because they do not support the necessary differencebetween sign types and sign tokens.

The point is more complex than it seems. The factthat not everything can be interpreted as a system ofabstract signs is a contested one. It is possible thateverything can be described or even explained byusing a system of abstract signs. However, it is usu-ally granted that such explanation is based on ide-alisation and abstraction in the sense that there is anontological difference between the explanation andthe explained: a biological theory of the apple cannotbe eaten. In the case of digital code, this ontologicaldifference is blurred by a common view according towhich an abstract specification of an algorithm (theexplanation) is in some sense more pure or real thanany instance of its implementation (the thing

explained). However, even in this case the ontologicaldifference prevails in that a digital device does notwork without the implementation. In the same senseas all linguistic systems border on or include inthemselves traces of their non-linguistic precondi-tions, all digital systems include traces of their phys-ical preconditions, and ultimately succumb to or areerased by something non-digital.6 Certain devices areable to function digitally under specific favourableconditions. These conditions do not have to degena-rate very much for the functioning to becomeunreliable with regard to the goals.

The existence of the digital is, then, doubly rela-tive. The digital exists with regard to something, inrelation to something; namely, with regard to somegoals and their reliable achievement as interpreted byus. This is a consequence of the fact that the digitaldoes not exist unless there is an abstract system ofsigns, which in its turn does not exist unless some-thing is interpreted as a system of signs. First, then,the digital is dependent on a system of signs, which inturn does not exist without interpretation. The tokenshave to be seen as tokens of a type. In this way theexistence of the digital is relative to the existence of alanguage and its interpretation, including all thebackground practices that a language implies. Sec-ond, the reliability of the read–write cycle is relativeto the goals set. Good examples are provided by thevarious forms and levels of error–correction andsample rate in digital compact disc technology.Another example could be the way in which countingwith ones fingers takes time to become reliable. Achild learning the skill has to practice hard in order toconsistently distinguish between the fingers and to beable to identify them. To oversimplify, a nine-pixelcopy and a nine million-pixel copy of Mona Lisa areboth digital copies of the painting, but the differencein terms of interpretation and recognition is huge.The necessary condition of reliability depends on thecontext.

Both of these ontological relativities are significantfor ethical questions. The first, the contextuality ofthe system of signs, is familiar from other fields suchas contemporary philosophy of meaning and mind.Nothing in the world as such forces us to regardsomething as a system of signs. Conversely, practi-cally anything can be read as a language, nothing canstop us from interpreting this or that as a system ofsigns.7 The existence of a system of signs and,

5 This is what Hubert L. Dreyfus implies, but does not spell out,

on p. 69 in Dreyfus (2001).

6 This is why Haugeland’s (1998) article is titled ‘‘Analog and

analog’’; in his view the digital is another type or subset of the

analog.7 John Searle (1980, 1992) is one of the prominent philosophers

of mind who has in various ways argued for the view that both the

existence of a syntax and of a semantics are relative.

DIGITALIGITAL NOMINALISMOMINALISM 227

consequently, of something digital, is therefore bothunder- and overdetermined by the world. Onlyhuman interpretation can resolve the matter. Thedigital, like the syntax and semantics of informationprocessing, are in the eye of the beholder.

The second relativity, the context-dependent reli-ability of the functioning of the system, characterisesthe fuzzy border between the analog and the digital.The required reliability and ‘‘perfectness’’ of thefunctioning are relative to the goal, e.g., that of lis-tening to music. A digital copy of, for instance, asong is never identical or perfect in all respects. It cannever in any sensible way reproduce the same vibra-tions of the same molecules. However, it can beperfect enough so that we are able to recognise andenjoy the song. The ‘‘perfect’’ digital copy is relativein the sense that no string of signs (sign tokens) canbe identical to another string of signs (sign tokens).The identity is on the level of idealisation: the stringsof signs may make possible the reliable enough rec-ognition of the underlying type, so that an illusion ofperfect match is created. Thus the functioning ofdigital devices is dependent on how well the devicescan sustain the human processes of abstraction andrecognition.

The existence of these relativities shows that‘‘digital being’’ includes a whole set of beliefs, a wholeworld-view, a whole ethos. The distinction betweentokens and types is powerful and has wide ramifica-tions. As is well known, long-winded philosophicaldebates can form around the question of if, how andwhere do abstract entities exist. The classical view-points of realism and nominalism readily spring tomind here, too. This philosophical side of the issuehas to do with questions of identity and difference.Some differences are abstracted away in order tocreate perceived identity. An illusion of sameness iscreated even though the physical world and experi-ence of it are in constant flux. Physical processes areunique in the sense that they cannot be repeated atwill. While digital processes can be repeated at will,they are in this sense non-unique.

With a certain rhetorical turn, it could be claimedthat ‘‘the digital’’, the belief in the existence ofsomething perfectly repeatable and media-indepen-dent that can be owned and operationalised, takespart in producing the truth of the existence of a dis-embodied and self-identical self (making things likeAI, transhumanism, genetic perfectibility facts of thepresent or the future). At the same time, this set ofbeliefs or forces, this Foucaultian episteme, shows itsown limits. Unique processes, those connected toparticular causal chains and those that do not lendthemselves to a separation between types and tokens,cannot, not even in principle, be digitalised. They are

not ‘‘code’’, at least not self-inclusive, formalisable,and calculable code. Keeping these relativities of theontology of the digital in mind, one can turn to theethical issues. What if the non-digitalisable processesinclude processes of knowledge that are vital toinformation societies? What if the appropriation ofthe digital into the scheme of private propertyrequires (because of the ontological nature of thedigital) social formations that are unnecessarilyclosed or oppressive?

Knowledge creation and ownership of digital code

Digital information technology provides the meansfor new types of sharing and distributing informationas well as the means for unprecedented forms ofcontrol and ownership of information. This is thetension included in all the digital utopias and dyst-opias. Let us take an example of this tension, anexample that crystallises the positions. One of themost interesting debates around the ownership ofinformation has been focused on computer software,not the least because software provides extremeexamples of both the proprietary and non-proprietaryforms of relating to code. On one hand we have theidea and the corresponding practices according towhich pieces of computer software are identifiableworks that are owned by their authors. Through thiswe get the proprietary system in which the ownersof software, typically software companies, give theuser the right to use the software under certaincondition specified in the end user license agree-ments. On the other hand there is the idea accordingto which computer software is algorithmic also inthe sense that it can not be owned: it is created inthe interaction of a body of individuals and organ-isations and is signed in the sense of ‘‘the authorwho discovered something’’ rather than in the sense‘‘the author who owns something’’. These extremesare exemplified, for instance, on the level of com-puter operating systems, where the dominantMicrosoft Windows system is the purest example ofa proprietary system, and the GNU/Linux systemmaybe the best-known example of a non-proprietaryand free system.

The main economic and social argument for theownership of software or code in general is that onlythe economic incentive of property and wealth canguarantee the prosperity of a society. The argumentsfor the freedom of software are similar to the argu-ments for the freedom of science: the speed ofdevelopment, the trustworthiness of the software andthe availability of code are all improved by the non-proprietary nature of software, and at the same time

TEREERE VADENADEN228

the values of co-operation and sharing are encour-aged.

The question can be investigated from the point ofview of knowledge intensive work and knowledgeproduction. Both the development of proprietarysoftware in software companies and the developmentof free software in volunteer ‘‘hacker’’ organisationsare certainly instances of knowledge intensive workresulting in goods that are digital information. Butthe contexts, including the underlying ethical, socialand political ideas, are quite different. The underlyingsets of beliefs in the different communities are notuniform, to be sure, but they do result in differentethical views.

Software development in a big software companyis institutionalised, more or less a Taylorist enterprisewith clearly defined and calculable subtasks leadingin a controlled manner to a coherent and well-definedresult. The knowledge production happens in anorganisation with its structure, aims and functions.Such knowledge production has been studied exten-sively, e.g., in view of the types of implicit and explicitknowledge circulating in the organisation.8 To use adefinition by the sociologist Scott Lash (2002) thefunctions of an organisation are structured throughnorms and the power legitimised by the norms. Anorganisation like this is hierarchical, with the ensuingchannels of command and division of labour.

In contrast to this, the development of free soft-ware of the GNU/Linux type happens in a widelydistributed and non-institutional manner in a globalvolunteer community co-operating mainly throughthe Internet. This kind of knowledge creation is notthat of the organisational type. Using Lash’s termi-nology, the free software community is a disorgani-sation based not on norms and rules but on sharedvalues. Lash illustrates the difference between anorganisation and a disorganisation with the differencebetween a church and a sect. The functioning of achurch is typically based on a certain hierarchy, a setof norms and legitimised use of power, whereas a sectis formed around a set of shared beliefs or convic-tions, visions, often exemplified by a charismaticleader. The difference is reflected in the operativeforce: in an organisation people are controlled andcommanded by the use of legitimate power, whereasin a disorganisation people can only be controlledand persuaded through the use of (physical or non-physical) violence. A disorganisation is formedaround a set of shared values, therefore the leader(the guru, in the case of Linux, Linus Torvalds) cancontrol the disorganisation by saying what is rightand what is wrong, what works and what doesn’t,

even though the leader does not have any legalpower. However, even if a disorganisation is non-institutional and informal, it is not by any meanschaotic. It can be much tighter controlled and ‘‘or-ganised’’ than an organisation, because the appro-priation of resources does not obey the normativisedcompartments of an organisation. A sect, a move-ment, or a tribe can function in a very effective andrational manner. This is one of the elements of freeand open source software development that has sur-prised and continues to surprise economic theory:according to received economic theory there is no realincentive for the volunteers to contribute to freesoftware development.

Against this background the question of owner-ship of intellectual property (in this case software)can be seen in detail. Software development in acompany, in an organisation, can benefit from thefact that pieces of information, knowledge and codeare seen as property. The ownership of code formsnot only an economic motivation, but also gives atool for the streamlining, rationalising and econo-mising of the organisational structure, gives a mea-sure for effectiveness and information flows, etc. Thepropriertarisation and commodification of intangible,abstract code is therefore a real asset to organisa-tional knowledge creation. This also creates anontological pull: it is good to believe in the indepen-dent existence of code. Knowledge creation inorganisations gains when code is appropriated interms of private property, when it is commodified andcontrolled through systems such as digital rightsmanagement.

In contrast, knowledge creation in a disorganisationis at the very least disturbed and at worst destroyed ifcode is the property of owners. In the GNU/Linuxcommunity, the developers take part out of their ownwill, they ‘‘scratch their own itch’’, and the systema-tised sharing of information guarantees that their workcontributes to a growing, common body of knowl-edge.9 The participants can contribute only in so far asthe flow of information is not controlled by ownership.The underlying values forming the community includea belief in the freedomof information and a passion forimproved information technology. A disorganisationis largely self-organising. The functions of the disor-ganisation are based on informal and non-institutionalco-operation and sharing, that cannot be compart-mentalised by decree or rationalised by use ofpower. The shared values are embedded in thefunctions of the disorganisation, and those functionsare at the same time the goal of the disorganisation.

8 See, e.g., Nonaka and Takeuchi (1995).

9 For views on open source and free software development see

DiBona et al. (1999) and Raymond (2001).

DIGITALIGITAL NOMINALISMOMINALISM 229

Disorganisations work as goals in themselves,therefore they do not, in general, benefit from theintroduction of external goals. The harm to knowl-edge creation in disorganisations is done if webelieve in the existence of code and assign aproprietary system of ownership for code.

From the point of view of the information society itis crucial to recognise the existence of these two dif-ferent types of knowledge work, and their differentpreconditions. A bias is produced by the fact that thevisibility of these two processes is different. Organisa-tional knowledge creation is organised also in the senseof having its own effective lobby and representativesystem, while disorganisational knowledge creationrelies on more underhand methods of social andpolitical influence. This bias is all themore unfortunateif there is reason to believe that information societiesare increasingly dependent on disorganisations. This isthe economico-political side of the issue that, however,cannot be separated from the ethical one. A societywhere knowledge is asymmetrically divided and frag-mented is less equal and democratic than one in whichknowledge is considered non-proprietary.10

Digital nominalism and the performative

contradiction of ownership of code

The example of free software, above, shows that thereare voluntary ways of using information technologyand ‘‘the digital’’ that deny that the ownership of‘‘intellectual property’’ would be essential for aninformation society. These ways of using informationtechnology can be further motivated by making areductio ad absurdum type of though experiment onthe idea that information and knowledge should beowned. Let us suppose that all knowledge is property,and that the use of that property is to be controlledby the owner of the property. Let us also suppose thatthe owners of knowledge want to share it only if thereis some gain for them in the sharing. Now, when achild is born into the society she does not know thatknowledge is owned. In order to learn that knowledgeis owned the child has to be taught a whole lot ofthings, and she has to live as a part of the society forextended periods of time. The education of the childis impossible if all knowledge is proprietary, unlessthe owners of knowledge want to ‘‘brainwash’’ thechild so that they can gain later on. Legally thiswould be impossible, however, as a newborn childcouldn’t enter into agreements that guarantee her

compliance after she has learned what ‘‘knowledge’’is, what ‘‘property’’ is, and the fact that all knowledgeis owned. If, however, such agreements are enteredonly in the mature age, the person can choose to stayignorant of the fact that all knowledge is owned. Shecan’t learn the fact, if the owners do not divulge theinformation, or she does not have to learn it, unlessshe asks from the owners. In all of the cases thepresuppositions lead into absurdities: the society inwhich all knowledge is owned is a closed one.

The main message of the admittedly naıvethought experiment is that knowledge first has to benon-proprietary for proprietary knowledge to bepossible. In the same way code first has to be non-digital for digital code to be possible.11 These arenot only temporal priorities but ontological facts. Ifthis is the case, it follows that the utopias of com-plete digitalisation and proprietary code necessarilydemand societies that are closed, because the inter-pretative system needed to uphold these utopias hasto include an ontological commitment to the exis-tence of abstract code. Abstract code is possiblebased on the affordances that concrete physicalthings and systems have under favourable condi-tions. One of the needed favourable conditions isthat people are willing to believe in code. Closedsocial formations are necessary, because an illusionof the reversion of the ontological priority (the ideathat the proprietary is more basic than the non-proprietary or the idea that digital code is morebasic than physical instantiation) can be createdonly through extensive social control. Finally,because an interpretative system is always a socialone, the belief in abstract, disembodied code mayhave a negative effect on our understanding of thevital functions of information societies.

Accordingly, we may suggest what types of socialpractices and structures of power circumvent not onlythe beliefs in the ‘‘uncontrollability’’ of (digitalinformation) technology but also point to ways ofusing technology that support equality and opennessin information societies. It can be assumed that as anontological stance digital nominalism, i.e., the beliefaccording to which code is always here and now inparticular concrete instances of functioning devices(including biological organisms), and according towhich ‘‘the digital’’ is only a nom-de-guerre for par-ticular types of electronic devices, directs our ethicaland political attention to the issues that are vital foran equalitarian information society. Moreover, it canbe assumed that the use of information and

10 There is also the possibility that a drastically asymmetric

division of knowledge produces a less effective and less innovative

society: see the argument and the references in Lessig (2001, 2004).

11 Furthermore, it can be claimed that organisations are based

on existing disorganisations: organisations are ways of channeling

the creativity of disorganisations; see Lash (2002).

TEREERE VADENADEN230

knowledge through the non-proprietary models ben-efits not only disorganisational knowledge creation,but also a sense of the (digital) technological interestsbeing malleable, not beyond all human control.

References

N. Block. Introduction: What is Functionalism? In: N.

Block, editor, Readings in Philosophy of Psychology.Harvard University Press, Cambridge, MA, 1980.

J. Derrida. Negotiations. Interventions and Interviews.

(edited, translated and with an introduction by ElizabethRottenberg) Stanford University Press, Stanford, 2002.

C.DiBona, S.Ockman andM.Stone, editors,OpenSources.

Voices from the Open Source Revolution. O’Reilly, Sebas-topol, 1999.

P. Drahos and J. Braithwaite, Information Feudalism. WhoOwns the Knowledge Economy? Earthscan, London, 2002.

H. Dreyfus, On the Internet. Routledge, London, 2001.J. Haugeland. Analog and Analog. In: Having Thought.Harvard University Press, Harvard, 1998.

S. Lash, Critique of Information. SAGE, London, 2002.L. Lessig, The Future of Ideas. The Fate of the Commons in a

Connected World. Random House, New York, 2001.L. Lessig. ‘‘Free culture’’ Presentation at the OSCON 2002Conference. Available: http://randomfoo.net/oscon/

2002/lessig/.L. Lessig, Free Culture. The Penguin Press, New York,2004.

D.M. Levy, Scrolling Forward. Arcade Publishing, New

York, 2001.I. Nonaka, Ikujiro and H. Takeuchi, The KnowledgeCreating Company. Oxford University Press, Oxford,

1995.E.S. Raymond, The Cathedral and The Bazaar. O’Reilly,Sebastobol, 2001.

J. Searle. Minds, Brains and Programs. Behavioral andBrain Sciences, 3: 1980.

J. Searle, The Rediscovery of the Mind. The MIT Press,Cambridge, MA, 1992.

R. Stallman. Free Software, Free Society. Selected Essays ofRichard M. Stallman. Joshua Gay, editor. The GNUPress, Boston, 2002.

DIGITALIGITAL NOMINALISMOMINALISM 231