Constructionism and its role in philosophy of information: An analysis of the relation between knowledge and reality in Luciano Floridi’s philosophy of information

(1)

Constructionism and its role in

philosophy of information:

An analysis of the relation between knowledge and reality in Luciano Floridi’s

philosophy of information

Abstract

In this thesis I approach the question of how knowledge and reality relate to each other from the perspective of constructionism in the philosophy of information (PI) of Luciano Floridi. The relation between knowledge and reality has been a philosophical debate throughout Western history. Floridi proposes that developing a philosophy of information is urgent considering far-reaching technological developments and that it facilitates an approach to knowledge and reality that offers a new insight in epistemology and ontology. I discuss his philosophy of information and the role of constructionism as an epistemological commitment therein. Furthermore, some of the critiques formulated against Floridi’s philosophy, most notably concerning his veridicality thesis, are placed in the context of PI and its theoretical background. Therefore, I argue that Floridi’s theory has to be appraised as a whole. This reading does not only shed light on the opposing arguments formulated in the literature, but it also appreciates PI as an informational ‘world view’. This entails the notion that the discussion of a theory, or parts of it, cannot and should not be decoupled from the context, with respect to the relation of knowledge and reality, in which it is formulated.

Sjoerd Drooger 10617116

Universiteit van Amsterdam Ma-Thesis Philosophy Thesis Supervisor: F. Russo Second Reader: F. Berto

(2)

1

Introduction

The relation between knowledge and reality is a topic that has occupied a central role throughout the history of Western Philosophy. From Platonic Idealism, to the Kantian transcendental revolution and the linguistic turn, major changes have occurred in philosophical thinking concerning both ‘knowledge’, ‘reality’ and the relation between the two. Recently, a new ‘revolution’ has been identified by the Italian-born philosopher Luciano Floridi, the ‘information revolution’. He describes a new form of self-understanding for the human race after the ‘fourth revolution’, which dethrones the special designation of ‘consciousness’ similar to the way Copernicus moved us out of the center of the universe, which constitutes the ‘first revolution’. The second moves the human race out of biological supremacy and into a line of evolution theory, credited to Charles Darwin, that retraces our origins to the ‘lesser’ animal kingdom. The third revolution that Floridi identifies is brought about by the work of Sigmund Freud, who theorizes that human consciousness is not as transparent and unproblematic as was once thought. The fourth revolution develops on par with technological developments in the second half of the 20th century, closely related to the rise of digital technology.1

In that same timespan, the philosophical concept of ‘information’ has entered the arena of knowledge and reality, although it is mostly a concept analyzed from a formal perspective, intertwined with information and computer science. Floridi characterizes the informational revolution as the dethroning of man’s unique role as information processors and aims to pay attention to the role the informational revolution plays in human self-understanding. Therein, he moves away from the more traditional approach in information sciences, where practical problems predominate scientific practice. It is safe to say that philosophical questions about data/information and the like have been largely silenced by practical question, questions of the form: how can we use and we improve upon (digital) technology? This is exactly what Floridi is aiming to address.

In order to do so, he has positioned himself and his project at an intersection of different disciplines; philosophy of technology, information sciences, computer science, ethics, formal epistemology and logic. This position is characterized as a ‘philosophy of information’, which Floridi presents as a new sub-discipline with its own sets of questions and methodology. The novelty of the sub-discipline is not constituted by the informational approach per se, therein Floridi relies on work done by Shannon & Weaver2_{, Bar-Hillel & Carnap}3_{and Fred Dretske}4_,

among others. The novelty, or rather the distinct nature of the discipline, is related to the large scale application of the informational approach. Floridi aims to use the notion of information to gather insight in the human understanding of the world and of itself in the post digital-revolution era. A simple intuition lies at the basis of this approach, namely that we have reformed our world through digital (informational) development to such an extent that traditional approaches to our relation to that world, and ourselves, do not suffices anymore. Hence the designation as information revolution.

1_{Floridi (2014)}

2_{Shannon & Weaver (1949)} 3_{Bar-Hillel & Carnap (1953)} 4_{Dretske (1981)}

(4)

3 Floridi defines the ‘philosophy of information’ (PI) as follows:

[Philosophy of Information is] the philosophical field concerned with (a) the critical investigation of the conceptual nature and basic principles of information, including its dynamics, utilization and sciences, and (b) the elaboration and application of information-theoretic and computational methodologies to philosophical problems.5

At a later moment Floridi expands on this definition when he incorporates the narrative of the four revolutions. Adams and Antonio de Moraes refer to this second step as moving from a new sub-discipline to understanding philosophy of information as an ‘ideological movement’6_{. They}

sum up Floridi’s new(er) view as “a replacement of a human-centric and conscious mind-centric world view with an information-centric view, in which ‘information’ acquires the central role”.7

Adams and de Moraes choose to focus on the first formulation, PI as a philosophical sub-discipline and disregard the ideological purpose they have identified. Other authors seem to either choose to do the same, or they are unaware of the ‘ideological’ component that Floridi’s fourth revolution entails.8

In this thesis I aim to do justice to the potential of the ‘ideological movement’, to take the claim that a human-centric world view is replaced by an information-centric one seriously. Therein, the status of the information-centric world view over the human-centric one is ontological. Adams and de Moraes point this out and B.R. Long draws attention to Floridi’s claim that if information cannot be understood as an independent ontological category, to what category it is reducible.9_{Furthermore, according to Long, Floridi posits the idea that informational ontology}

can be understood as the basis to which material ontology can be reduced.10_{These are rather}

controversial claims; I will discuss these further in chapter three.

Generally, Floridi’s work has been received positively, especially his project of establishing philosophy of information as a mature sub-discipline. This is not to say, however, that the commentators are in agreement with respect to which part of Floridi’s work deserves most attention. Luckily so, since it ranges from ethical theorizing about the post digital revolution world, to formal argumentation concerning the way Floridi aims to solve the Bar-Hillel-Carnap paradox in information theory. In general, commentators of Floridi can be split up in two camps. Firstly, there are those that deal with Floridi’s ethical theory, which is postulated with a specific post-digital-revolution world in mind. In the information age that Floridi envisions, human beings share their role as information processors with ‘artificial companions’ that we may be unable to distinguish from a regular companion. An ethical theory that deals with a situation where an entity that more traditional ethics would designate as a mere ‘thing’ becomes part of the same sphere as us, will have to find a more inclusive way of viewing the world. This is exactly what Floridi aims to do, but his moral theorizing is not without critique.11_{Although this}

strand of the debate is interesting, it falls outside the scope of this thesis and thus it is not discussed further.

5_{Floridi (2002), p. 137; also in Floridi (2011a), p. 14} 6_{Adams & de Moraes (2014), p. 161}

7_Ibid.

8_{see for example: Brenner (2010); Fyffe (2015)} 9_{Long (2014), Floridi (2010)}

10_{Long (2014)}

(5)

4 The second camp that can be distinguished is one that takes a formal epistemological approach, on par with the philosophy of information as it has been practiced in the past. Floridi proposes a theory of information that combines elements he derives from past work in the field of information theory, from Bar-Hillel and Carnap, and Dretske, among others. He adds ideas he derives from more recent developments in computer science. Floridi’s personal contribution comes in the form of the veridicality thesis, which finalizes the core of his view on information: σ (a infon12_{) is an instance of semantic}13_{information if and only if it consists of well-formed,}

meaningful and truthful data. The critiques in this camp focus on this formal definition and go into detail as to which problems they conceive with the theory. Some of the critique in this camp can be characterized as disagreement (Long 2014) or even flat-out rejection (Adriaans 2010), while others aim to mark potential problems within the theory (Lundgren 2014, 2017).

In this thesis I engage with the topics that relate to the second camp of commentators. My aim is to shed light on the importance of Floridi’s ‘ideological movement’, mentioned above, and relate it to some of the topics of discussion concerning PI, most notably the veridicality thesis. That being said, the goal of this thesis is not to propose some decisive reading of PI that will do away with any and all critique, problems or controversies. Neither does it aim at rejecting the comments and opposing views that have been formulated over the years. Ideally, some of these comments will be put in a different light, which does not mean they will be resolved per se. In order to reach this goal, one of the underappreciated14_{elements of Floridi’s work will be}

emphasized, namely his theory of knowledge being constructionist.15_{Floridi explicitly rejects a}

mimetic understanding of knowledge and adopts a Kantian inspired approach instead. This entails that an epistemic agent does not have access to the world ‘out there’ in an unproblematic sense, which is not to say that the reality independent of the epistemic agent is denied altogether. It means that access to the world, knowledge if you will, is at all times ‘coloured’, in Kantian flavor, by the epistemic glasses the agent wears. This is not to say that there is some unlimited amount of potential understanding of reality. In Floridian terminology ‘data constrain our theories even if they do not determine them’16. Constructionism walks a fine line in between

realism and constructivism17_{. It can be hard to keep track of in its own right, let alone not losing}

sight of it in a dense, technical and generally formal work of philosophy as PI. However, it is worth emphasizing as a theory of knowledge in Floridi’s work as it plays a crucial role in understanding the overall theory.

The aim of the thesis can thus be described as exploring the relation between the ‘background issues’ concerning the philosophy of information and the theory itself. It does not constitute an exhaustive analysis of this field that would be too extensive, instead it focuses on taking constructionism as an entry point and relating that to key notions in PI, namely truth and knowledge. In doing so, the analysis leaves several topics unattended, but this is a necessity

12_{the entity that constitutes a piece of information}

13_{In this thesis I accept the understanding of information as being semantic, i.e. consisting of meaningful} data. Understanding information semantically is not necessarily accepted. Adriaans (2010) for instance, argues that information is best understood more formally, mathematically. The discussion concerning the ‘semantic’ nature of information is noted, but side-lined in this thesis.

14_{It seems that most commentators that criticize the formal aspects of Floridi’s theory (Adriaans, Long,} Lundgren) do not incorporate this crucial part of the theory as a whole.

15_{Floridi (2011b )} 16_{Fyffe (2015), p. 282}

17_{Floridi presents his constructionist epistemology as a position in between realism and constructivism.} These terms are discussed in chapter 1.1

(6)

5 regarding the scope of the project. In the analysis between constructionism and PI, the larger theme of knowledge vs reality is addressed through the perspective of constructionism as well. As such it aims to contribute not only to understanding the philosophical work of Luciano Floridi, but also to understanding the constructionist position more generally and its relation, as an epistemological approach, to ontology.

The thesis is structured as follows: in chapter one I will first introduce constructionism as an epistemic position, in the context of Floridi’s analysis of the user’s knowledge vs maker’s knowledge tradition. In chapter two, constructionism is postulated in opposition to the traditional epistemological view, which is related to the user knowledge position, namely the Justified True Belief model (JTB). In that analysis I discuss Floridi’s view on the shortcomings of JTB and his alternative approach in the form of the Network theory of account model (NTA). The main reason for doing so is to inquire whether Floridi is successful in opposing JTB, without assuming the constructionist position beforehand. I will show that although Floridi walks a fine line, he can be defended to do so. Therefore, the alternative that he proposes, NTA on this specific issue but the informational analysis in combination with a constructionist epistemology on a large scale, is posted as an alternative view, moving past the traditional epistemological model in the form of JTB. In chapter three, I discuss the relation between information and knowledge, by imposing the veridicality debate concerning information, and Floridi’s ontological commitment, namely Informational Structural Realism (ISR). Finally, I argue that the philosophy of information is fundamentally grounded on the constructionist position and that, furthermore, the constructionist position is directly related to the ontological commitment formulated in ISR. I conclude that discussions concerning specific points in the philosophy of information should be contextualized in the constructionist model, in order to grasp the full force of the revolutionary ‘ideological move’ that Floridi proposes.

(7)

6

1. Constructionism

In this chapter I discuss Floridi’s constructionist position in terms of his historical narrative of the dichotomy between the maker’s and user’s knowledge traditions. First, I go into the idea of ‘user knowledge’ and discuss Floridi argument for its hegemony, which he calls the ‘Platonic Dogma’. Secondly, I discuss the constructionist position as a form of maker’s knowledge.

1.1 Maker’s vs User’s Knowledge

1.1.1 The user-knowledge tradition

Alongside the historical narrative that cumulates in the fourth revolution, Floridi also identifies a tendency that stretches from the ancient Greeks to present day.18_{This tendency entails}

understanding the position of the epistemic agent in the knowledge production process as a ‘user’, rather than a ‘producer’. In terms of the relationship between knowledge and world, we humans passively conceive the world and produce ‘mimetic’ representation, that is: knowledge, of that world. The emphasis lies on the notion of passivity when understanding the human position as ‘user’. The world, out there, is a vast and static entity for us to ‘consume’ or ‘discover’. Passivity, therein, means that the epistemic agent does not contribute to the world in its knowledge production process, but merely represents it.

According to Floridi, this way of approaching the relation between knowledge and the world originates in the work of Plato. Floridi’s reading of Plato19_{entails differentiating between using,}

making and imitating an object. This is applicable to any sort of object. In the case of a chair, there are those who sit on it, those who build it and those that paint a picture of it. Another example is that of ‘reins and a bit’. The user is the horseman, who employs the tool to control the movement of the horse. The maker is the smith who produces the reins and bit, but does not use them himself. The imitator, again, is someone who depicts the object in some way. Plato employs this distinction to argue against the ‘imitators’, or the ‘sophists’ as is generally accepted. However, what interests Floridi in this discussion is the contrast between maker and user.

Plato employs a distinction in types of knowledge in order to grasp the user vs. maker distinction. The two types of knowledge are techne and episteme respectively. Episteme is commonly understood as ‘theoretical knowledge’, while techne refers to ‘practical knowledge’ related to a ‘craft’ or ‘art’. 20_{The difference between the two types of knowledge is captured}

most clearly if we consider an object of which both making it and using it require some expertise, as the reins and bit but not the chair. Another example is that of a musical instrument. Making a violin requires knowledge, but playing a violin requires a different set of knowledge. With the violin the questionability of the Plato’s move is most apparent, because making the violin is taken as knowledge in the form of techne and knowing how to play the violin as episteme. The violin builder is thus understood as having practical knowledge while the player has theoretical knowledge. There is something counterintuitive here as, at least in our contemporary context, the knowledge to play an instrument is often, at least in part, seen as a skill, as something practical. After all, one cannot learn to play the violin from theoretical reflection alone, for a large part it requires practice.

18_{Floridi (2011b)}

19_{Ibid., pp. 286-290} 20_{Parry (2014)}

(8)

7 Plato’s move to the user’s knowledge perspective, the Platonic Dogma21_{as coined by Floridi, has}

to be understood in light of his philosophical framework. According to Plato all the objects that make up the world as we conceive it are reflections of the ideas that occupy a world other than ours. The world of ideas is a divine sphere where all objects as we know them originate in a sort of blueprint to object relationship. So there is the ‘idea’ violin, the ‘idea’ of a chair, the ‘idea’ rabbit etc. Every object in ‘our’ world is a reflection of the divine original, the ‘idea’, which is created by the Ur-maker, God. Insofar an object is a reflection of an original idea, it is also, in a sense, an imitation.22_{The position of a maker in the Platonic context is thus not that of the}

original maker, since that role is always reserved for the Divine. In that sense, someone who ‘makes’ knowledge, for instance geometers and astronomers according to Plato, are merely communicating knowledge that was, in a way, already there.23

Although the Platonic conception of the relation between knowledge and world, and the role of the epistemic agent therein, may seem alien in our contemporary understanding, the role of Plato’s work should not be underestimated. This seems to be the message that Floridi wants to convey. Of course Floridi is not alone in pointing out the importance and influence of Plato’s philosophical work, as he is widely regarded as one of the most influential philosophers of all times. However, Floridi is arguing that in the Platonic Dogma we find the origin of the user’s knowledge perspective that has been the prominent way of understanding the role of the epistemic agent in the relation between knowledge and the world, all the way up to present day. So the question, at this point, is: how does the user’s knowledge perspective manifest itself in the present day, especially with respect to contemporary views on knowledge production?

1.1.2 Commonsensical realism

Floridi relates two key notions to the user-oriented approach to human knowledge production, namely passivity and mimesis. These notions are also related to a common approach in epistemology, that is naive or commonsensical realism. This is a traditional approach to the question of how knowledge relates to the world. In the commonsensical approach, the world is taken as a collection of material objects. Our access to that world is usually understood as direct and unproblematic, through our sense-experience. Knowledge consists of mimetic representations of the world, that we produce based on our sense-experience of that world. The key is that our relation to the world is a passive one, knowledge is based on passive reception of the fact that we call reality. The sort of knowledge that is produced by such an approach usually takes up the form of declarative sentences, in other words propositional knowledge: “It is the case that X”. How do we know that X is the case? We have identified it such and such a way, i.e. through empirical research. Note that Floridi connects this form of ‘naive realism’ to his understanding of the Platonic Dogma without further explicating why these two ideas are related. His approach is to emphasize the potential supplementary nature of the combination of user-knowledge and naive realism.24_{In this thesis, I follow that approach.}

I propose another way of understanding the commonsensical approach to knowledge, namely through the notion of ‘knowledge-that’ as opposed to ‘knowledge-how / why’. The importance of ‘knowledge-how’ in contrast to ‘knowledge-that’ has been emphasized by Hans Radder25_{. He}

21_{Floridi (2011b), p. 282} 22_{Plato (1989), Republic, 601c} 23_{Floridi (2011b), p.289} 24_{Ibid., p. 283} 25_{Radder (2017)}

(9)

8 believes that ‘knowledge-how’ has been receiving too little attention, especially in the context of justification of knowledge. Justifying a knowledge claim can take two forms, according to Radder. On the one hand there is weak justification, which is based on authoritative testimony. On the other hand, strong justification is based on the epistemic agent being able to provide the justification themselves. This distinction follows the know-that/know-how divide, in the sense that knowing that something is the case can rely on authoritative testimony. Knowing how something is the case already implies that you are able to provide a complete account of how X works. Thus, you are presumed to be able to justify the knowledge claim. Radder’s notion of knowledge is strongly connected to the justification of it. In a user oriented approach to knowledge as Floridi characterizes it, stressing the importance of justification clashes with the ‘passive’ stance that the epistemic agent is supposed to have. Even more so if we follow Radder’s emphasis on justification as a process, an activity that the agent has to actively engage with. A similar tension arises when we think of commonsensical realism in relation to the justification of knowledge.

Commonsensical realism is fundamentally intertwined with a propositional account of knowledge. In Radder’s terms, we could say that the authoritative testimony is sufficient because the sense-experience of the world itself is unproblematic. In a way, our experience of the world functions as an authoritative testimony. In contrast with Radder this does not constitute a weak justification, quite the opposite; it is the strongest form of justification, as there is no reason to doubt it. After all, the authority is reality itself and not, as is implied in Radder’s framework, some other epistemic agent (testimony) which has some authority but is not omnipotent. Understanding the relation between knowledge and world in this commonsensical way has not been without criticism. One of the ways to attack this approach is by questioning the unproblematic status of our sense-experience. This results in a position known as scepticism. Sceptical arguments come in a variety of degrees and can go as far as claiming that our experience of reality is, in its entirety, a human construct. Such a position is denoted as constructivism, referring to the role that the human mind is designated in such arguments.26

So, the user’s knowledge tradition has, according to Floridi, played an important role throughout history and results in a commonsensical realist approach to the relation between knowledge and the world. Floridi identifies a historical problem in the user perspective on knowledge which he has called Plato’s Dogma. Recall that Plato emphasizes the user position of knowledge having ‘more’ or ‘better’ knowledge than the maker, a move that can be ascribed to the role of Platonic Idealism. Floridi claims that this move was not necessary and that it has left us with a problem. Namely, Plato’s Dogma and the user’s knowledge tradition presuppose a ‘form of external realism’ and ‘a theory explaining our potentially noiseless access to it’.27

Furthermore, criticising the commonsensical realism approach to knowledge easily falls prey to a sceptical argument that turns the notion on its head. Instead of external realism and an unproblematic relation to it, the sceptical argument denies the existence of an external reality and claims a constructivist relation between the epistemic agent and the world that he experiences.

26_{Floridi (2011b), pp. 284-285} 27_{Ibid., p. 290}

(10)

9 Floridi employs his analysis of the origin of the user oriented knowledge tradition and points to the possibility of taking up a maker’s oriented approach. This is not to say that there has not been given any thought to the maker’s knowledge approach until the informational revolution. Floridi identifies, following Pérez-Ramos28_{, the work of Francis Bacon as a turning point, when}

Bacon states “To know truly is to know through causes”29_{. Here the epistemic agent is appointed}

a maker’s role with respect to knowledge production. As such, the maker’s knowledge tradition can itself be traced historically as a movement that has, at least since the scientific revolution, competed with the user oriented approach. However, at this point I will not go further into the historical narrative that could, and perhaps should, be researched with respect to the maker’s knowledge tradition. With the aim of this thesis in mind, I will accept Floridi’s reading of Plato’s Dogma as resulting in user oriented commonsensical realism. Furthermore, in order to explore the potential role that a maker’s knowledge perspective can play, especially in the context of the informational revolution, I will turn towards Floridi’s proposed alternative to common sense realism and its sceptical counterpart, namely constructionism.

1.2 Constructionism as the maker’s approach to knowledge

Floridi proposes to take a ‘radically moderate’ stance30_{with respect to the form of realism that}

logically follows from the user oriented knowledge tradition. This radical moderate approach points towards the pitfall that Floridi identifies in understanding the choice that any epistemologist has to make between, either supporting a naive notion of external realism with no constructive role for the epistemic agent, or denying the existence of an external reality altogether and adhering to a totally constructivist approach. In other words, he proposes a position in between realism and constructivism, with elements of both approaches. Furthermore, in order to dispel the false dilemma, he points towards the role of the user knowledge tradition. Understanding the role of the epistemic agent and the relation between knowledge and reality in passive and mimetic terms offers the notion that agent and world are completely independent from each other in the knowledge production process. However, if we move towards a more maker oriented approach to knowledge, the different realms, agent, knowledge and world, are understood in an interrelated and interdependent manner.

Floridi refers to the position that he endorses, situated between realism and relativism/scepticism, as constructionism. This position aims to retain the constructing element in constructivism, without denying the existence of an ontologically independent reality ‘out there’. It entails understanding the position of the epistemic agent as the maker of knowledge, rather than the user. This means that, at least with respect to the production of knowledge, the world is not understood as something independent that can be passively conceived. However, it does not mean that the world is understood as a mind-dependent realm. That would be constructivism. Floridi acknowledges the delicate balance that this position involves, he evokes the metaphor of pulling a tablecloth from under a fragile set of dinner service. The aim is to change the foundation that all knowledge that we have accumulated over the past two millennia is built upon, without disturbing the complex network that is constituted by that knowledge.31

28_{Péres-Ramos (1988)}

29_{Floridi (2011b), p. 291} 30_{Ibid., p. 285}

(11)

10 The first step to reach this goal is to displace the foundation that is in place. Floridi has attempted to do this by framing said foundation as a user knowledge approach and argues that this is an unnecessary and deficient position. The second step is to offer a viable alternative that does not negatively influence the results that have been achieved under the original approach. Floridi’s proposal is to adopt a maker’s knowledge approach, in the form of constructionism. This position reframes the role of the epistemic agent in the knowledge production process without falling into the relativistic trap of a sceptical approach to reality. Whether or not constructionism is successful in the task of not deconstructing the building of knowledge that has been built on the former foundation remains to be seen. Such a question is of too large a scope to be treated in this thesis. The question that can be answered is whether constructionism can, in theory, function as epistemological commitment and, furthermore, what it would entail ontologically and what it means for the relation between those spheres, between knowledge and world.

With respect to moving from a user- to a maker-oriented approach, a large part of the work has been done above. Following Floridi’s reading of Plato’s Dogma, the position of a user knowledge tradition has been weakened. The question now is what a maker’s knowledge perspective entails. In the maker’s tradition, knowledge is not out there, ready to be conceived by us, but we have to engage with the world and construct our understanding of it in the process. Floridi states: “Knowledge is not about getting the message from the world; it is first and foremost about negotiating the right sort of communication with it”32_{. Going back to the platonic}

framework that was identified as the basis of the user oriented approach, the maker’s perspective has one clear obstacle to overcome. In Plato’s context, the original, the real, maker’s role was reserved for the Ur-maker, for God. Ascribing it to the epistemic agent would, in that context, result in a blasphemous position. Nowadays this is not so much the problem, however, besides the religious component, the Divine position of Maker also accounts for the creation of the ‘world’-side of the knowledge-world relation. Surely, the maker’s knowledge approach does not entail the claim that the world has been created by humankind. In order to deal with this tension we have to invoke the maker’s perspective as constructionism.

1.2.1 Roots of constructionism

Floridi does not go into detail regarding the origins of the constructionist position. He does mention that his form of ‘philosophical constructionism’ goes hand in hand with constructionist learning theory developed by Papert and Harel.33_{In order to gain insight in the origin of}

Floridi’s constructionist position, I will discuss the constructionist learning theory here and relate it to Floridi’s broad epistemological use of the constructionst notion.

Papert and Harel’s notion of constructionism originates from educational theory. In a general sense it can be understood as thinking of the process of learning as ‘learning-by-making’. It shares the element of constructivism that learning entails the building of knowledge structures and it adds a social component in the sense that understanding learning as construction is especially successfully if the construct is a ‘public entity’. This means that the construct is not a solipsistic object, but has to bear upon a form of intersubjectivity. Papert and Harel do not see a fundamental problem in presupposing a form of intersubjectivity, although they do not address the issue specifically. From the perspective of an educational theory, it can be defended that the

32_{Floridi (2011b), p.284} 33_{Ibid., Papert & Harel (1991)}

(12)

11 issue is not pressing as the status of intersubjectivity is not at stake. In terms of education and modes of learning/teaching we can rely on a traditional realist position. However, Floridi cannot evade the issue, since he uses it to oppose the traditional realist position. His method of dealing with this issue is expounded upon in the next section.

Papert and Harel identify ‘instructional’ education as a traditional approach that lacks resonance with many students in the educational system. Constructionism is not a ‘hard’ alternative according to them, in the sense that they do not offer a knock-down argument to the traditional approach, and neither do they offer a definitive line of reasoning that ends up in constructionism. Rather, constructionism appeals to a form of ‘resonance’34_{, where one comes to}

see constructionism, in a more or less intuitive way, as an alternative to instructionism. It is, therein, simply a way of looking at how people think, how they learn. Even if, in science, there was a claim on what is the ‘best way’ to think about thinking, there would still be those who prefer their own way rather than the ‘best way’.35

This line of reasoning leads Papert and Harel to distinguish two kinds of claims about constructionism. First, a weak claim, that constructionism suits some people better than other (i.e. instructionist) ways of learning that are being employed nowadays. Secondly, a strong claim, that constructionism as a mode of teaching/learning is better for everyone, because it allows for a much wider variety of ‘intellectual styles’ to exist simultaneously. The latter claim is stong, especially in the sense that a constructionist approach to learning supposedly allows for other modes in be incorporated. In other words, constructionism is better than instructionism because it allows for instructionism as well as other approaches. Rather than arguing in favour of this claim, Papert and Harel point out that potential is there.36_{Constructionism, in their view,}

has not reached the level of maturity and stability that is needed for treating it in the same way as the traditionally accepted approach. In its current state, at the time of writing, they refer to constructionism as ‘pre-paradigmatic’.37

1.2.2 How Floridi employs constructionism

Note that Papert and Harel’s view on constructionism as a potential alternative, in educational theory, to the traditionally accepted approach, in the form of instructionism, corresponds directly with Floridi’s use of the notion as an alternative, in epistemology, to the traditionally accepted notion of commonsensical realism. Moreover, where Papert and Harel rely on practical examples to support the claim on constructionism, Floridi adds a theoretical argument based on his historical narrative concerning the user vs. maker oriented tradition of knowledge. In this way Floridi develops constructionism further by transposing it from the context of educational theory to the field of epistemology. Therein he attempts to provide an argument of where the divide that Papert and Harel identify as constructionism vs. instructionism can be found in epistemological thought. In that sense, Floridi situates the question originally raised in terms of educational theory in the general debate of the relation of knowledge and reality and the way the epistemic agents is posited in that relation. Based on the understanding of the status of constructionism, as ‘pre-paradigmatic’, the potential role of constructionism in the epistemological debate is not a full contender of the traditional approach as identified by Floridi. However, if Floridi is successful in providing a theoretical basis for the maker’s knowledge

34_{Papert & Harel (1991), p. 3} 35_Ibid.

36_Ibid. 37_Ibid.

(13)

12 approach, in the form of constructionism, as opposed to the traditional user oriented view, than he has made a major step in putting constructionism on par with the traditional approach. There are two parts to the argument in favour of taking Floridi to be successful in this task. Firstly, there is a historical component in which Floridi shows that the maker’s knowledge tradition is not without support in philosophical thought. The role of Francis Bacon has already been alluded to. Bacon’s work is characterized as a turning point, accumulating out of an Aristotelian-Scholastic tradition where the notion of knowledge is already connected with ‘understanding of the causes that bring about the known’38_{. Bacon, and the scientific revolution}

more generally, are said to breach the notion that the role of the epistemic agent was understood as a passive one, with respect to its relation to the world. Floridi writes that with Bacon and the scientific revolution “both the philosophy of language and the philosophy of mind should be at least equally concerned with creative rather than reproductive capacities”39_{. Other}

advocates of a maker’s knowledge tradition can be found in Giambattista Vico, Thomas Hobbes and John Locke. Vico is especially known for his emphasis on the role of the epistemic agent as a maker of the knowledge he possesses. Verum esse ipsum factum, what is true is what is made, is his slogan. Interestingly enough, Vico identified the constructing role as related to what is nowadays referred to as the Humanities, whereas Bacon relates the notion more closely to the ‘scientific’ realm.

It should be noted that there have been more negatively coloured conclusion with respect to the history of the maker’s knowledge tradition. Jaakko Hintikka40_{points towards the sceptical,}

‘constructivist’, component in the maker’s approach. The sceptical component has been identified as the potential of placing complete constructive power in the human experience, which means that reality is understood as a product of human thinking/understanding. Hintikka claims that this ‘pessimistic’ conclusion was used in order to “extol the superiority of the Divine practical reason over the human one”41_{. In other words, the sceptical component in the maker’s}

knowledge approach gave rise to the possibility of employing this scepticism, to discredit the approach by pointing out it’s sceptical nature, and re-establish the superiority of the Divine, as true maker. Floridi identifies a deficiency in this conclusion and formulates the, according to him, fundamental task of modern constructionism:

that of soldering together the Platonic dichotomy between human making and divine making and hence between the two different epistemological ‘arts’, by reinterpreting the dualism between human and divine knowledge in ontological terms, between noumena and phenomena, between the reality of data in themselves and the world of information as we know it.42

This brings us to the second part of Floridi’s task of establishing the basis for a maker’s knowledge approach. Firstly though, note that the historical argument in the sense of pointing out the prior existence of a maker’s knowledge approach does not contribute to establishing the legitimacy of the position. What it does is placing the argument in favour of a maker’s perspective in a historical trend. In order to make use of this, the potential opposition related to 38_{Floridi (2011b), p. 290} 39_{Ibid., p. 291} 40_{Hintikka (1974)} 41_{Ibid., pp. 85-86} 42_{Floridi (2011b), p. 292}

(14)

13 the historical character of the approach must be dealt with. This is what Floridi aims for in the quote presented above. Modern constructionism, as a continuation of older forms that were susceptible to scepticism and pessimism, has a fundamental task. This task is described as ‘reinterpreting the dualism between human and divine knowledge in ontological terms’. Thereafter, Floridi addresses this dualism as between noumena and phenomena. This Kantian distinction is presented without introduction. Floridi employs the notions to ‘reinterpret’ the dichotomy, as he identified in Plato’s work, between human and divine making, or human and divine knowledge, ontologically. Divine knowledge is placed in the sphere of noumena and human knowledge in the sphere of phenomena.

Recall that in Kant’s philosophy, the phenomena are the objects of knowledge as experienced by the human mind. These objects are understood as based on a ‘real’ object, independent of the human cognitive condition, the Ding an sich, who reside in the realm of the noumena.43_Floridi

employs this distinction to reconcile the human role of makers of knowledge in the sense of phenomena and the divine role as maker of noumena. With that he claims to place his notion of the maker’s knowledge approach, his constructionism, in between the divide of realism and constructivism. He evades the sceptical pitfall by incorporating the main element of realism, that of an external reality, in the sense of the noumenal realm, ascribed to divine construction. The access that the human mind, the epistemic agent has to this realm is, however, far from unproblematic. In our knowledge production we are restricted to the realm of phenomena. We are the makers of knowledge and that knowledge has its own status. It relates to the experience of the world as ‘we’, the human mind, have it. This does not mean that there is no reality apart from the human experience. At the most it entails that said reality cannot be accessed, ‘touched’, or approached, outside our experience of it. This boils down to a theoretical presupposition that the world, ‘out there’, exists independently of our knowledge of it. Therefore, the notion of an intersubjectively shared reality is accounted for; recall the issue that was pointed out concerning constructionism in Papert’s context of educational theory.

Potentially, this still gives rise to a problem, since at least with respect to constructionism it is not particularly clear how the phenomena relate to the noumena. Floridi takes up the view that it is not productive to deny the existence of reality44_{and therefore a theoretical solution for}

distinguishing the realm of knowledge from the realm of reality without making phenomena independent of noumena should be enough. The critic could, however, demand an explanation of the workings of relation between noumena and phenomena in constructionism. We could look for such an explanation in the work of Kant, since the theoretical move is directly related to his work. However, Floridi does not go into such an analysis. Instead he develops a different line of reasoning, one that has already been part of his treatise of the traditional user’s knowledge perspective. I will come back to the relation between the phenomena and the noumena in chapter three, when I discuss Floridi’s ontological commitment. There, I present an understanding of the relation between objects of knowledge and ontological objects in the context of Floridi’s constructionism.

43_{Kant (2007), Longuenesse (2000)} 44_{Floridi (2011b), pp. 284-285}

(15)

14

1.2.3 Constructionism as methodological and meta-philosophical

This chapter has mostly dealt with constructionism as the maker’s approach to knowledge. The previous section ended with the conclusion that the maker’s knowledge approach to knowledge is posed as an alternative to traditional, ‘user’, views on knowledge. This alternative is connected to the notion of constructionism as a pre-paradigmatic educational theory. I invoked Floridi’s argumentation concerning the theoretical basis of his maker’s approach and claimed that he is, therein, attempting to put his view on equal footing with the traditional, user-oriented one. I concluded that he is not completely successful in that regard, as we have traced the argument to a Kantian move, where ‘Divine’ making is placed in the noumenal realm and ‘human’ making in the phenomenal realm. I stated that Floridi sees no salvation in denying reality altogether. This raises two problems, firstly, why is it not productive to deny reality altogether? Secondly, if we accept the claim that denying reality is not the way to go, how do the noumena and phenomena interrelate?

In his proposal of constructionism, Floridi does not offer answers to these questions. Rather, he proposes a distinction between the meta-philosophical level and the methodological level of constructionism. This chapter, so far, has dealt with the meta-philosophical level, in that it presented constructionism as a view on knowledge. Methodologically, Floridi proposes, one does not actually have to adhere to the maker’s knowledge approach in order to work with constructionism.45_{Intuitively, this seems like a big gap. Why would anyone want to use the}

constructionist method, if they are not convinced by constructionism on a meta-theoretical scale? I propose to understand this move in light of the overall tendency regarding the proposal of constructionism as an ‘alternative’. Floridi moves from the meta-theoretical level to the methodological level, while leaving room for those who do not wish to adhere to his meta-theoretical alternative. He, therefore, carves out his ‘constructionist’ methodology as made up of three parts: minimalism, the method of levels of abstraction, and constructionism. Cleary, the first two are supposed to be disconnected from constructionism (methodologically speaking) to such an extent that they can be regarded separately. Apart from the question of how different pieces of a philosophical method can be related and at the same time can be considered separately, the question rises of how the methodological level (with or without the constructionist component) relates to the meta-theoretical level. Floridi states: “you might wish to help yourself with the latter two tools [minimalism and method of levels of abstraction] and stop short of accepting a maker’s knowledge perspective”46_{. This indicates that the method can}

also be taken separately from the meta-theoretical view, in the form of the maker’s knowledge approach.

In the remainder of this thesis I aim to shed light on the difficulties that raise with the claim that method and meta-theory can be separated. Instead of side-lining the constructionist position, I develop it further, relating it to different parts of Floridi’s philosophy of information, in which his ‘constructionist’ methodology plays a crucial part.47_{The scope of the thesis does not allow}

for a detailed discussion regarding the constructionist method. Instead, I present the most

45_{Floridi (2011b), p. 294} 46_Ibid.

47_{Floridi states that in his work The Philosophy of Information he attempts to combine his reading of} Kant’s epistemological project, “how empirical knowledge .. develops through the constructions of our information about the world”, with a constructionist philosophy of information. (Floridi, 2011b, p.293)

(16)

15 important aspect, the method of levels of abstraction, here and refer to Floridi’s own work48_for

more details. Furthermore, I re-address the potential issues identified on the meta-theoretical level in the conclusion of the thesis.

1.2.4 Method of Levels of Abstraction

Going back to Floridi’s definition of PI, there are more extensive characteristics that can be added, however for this thesis the reading formulated by Brenner will be followed. He points towards two additions, firstly to the notion of the dynamics of information. There he states “[the dynamics of information is] the constitution and modelling of information environments, including their systemic properties, forms of interaction, internal developments, etc.”49_.

Secondly, he draws attention to the importance of the method of levels of abstraction (LoA) in Floridi’s philosophy of information. A LoA is defined as a “finite but non-empty set of observables, where an observable is just an interpreted typed variable, that is, a typed variable together with a statement of what feature of the system under consideration it stands for”50_{. This}

is a rather formal definition, which makes sense as Floridi has, in part, borrowed from the field of computer science where discrete mathematics is employed to analyse information systems. However, as is often the case in Floridi’s work, he also offers a more intuitive definition of his method. He invites us to think about a conversation between Anne (A), Ben (B) and Carole (C).51

Anne is a collector and potential buyer, Ben tinkers in his spare time and Carole is an economist. They voice the following statements:

a. Anne observes that there is an anti-theft device installed, it is kept garaged when not in use, and has had only a single owner.

b. Ben observes that its engine is not the original one, that its body was recently repainted, but that all leather parts are very worn.

c. Carole observes that the old engine consumes too much, that it has a stable market value, but that its spare parts are expensive.

Note that at this point we do not know what the object of the conversation is. Each view is expressed according to certain interest. These interests are teleological driven, in the sense that they are voiced in accordance with a certain goal/context, a buyer, a mechanic and an economist. Each of these perspectives gives voice to a level of abstraction. Without knowing what the referent is, it could be a car, a plane, a motorcycle, it is a source of information, which can be understood as a system. Each LoA stands for a possible analysis of that system. These three LoA’s do not exhaust the potential list of different analysis that could be made of the system in question. Floridi proposes to understand each analysis of the system, each LoA, as producing a model of the system. This treatise leads Floridi to offer the more formal definition that Brenner reproduced. A final attempt to clarify the LoA is made by Floridi, he compares a LoA to an interface, or rather an interface to a collection of LoA’s. An interface is situated in between two systems and relates the output of the one system into the input of the other and the other way around. So in the example of the conversation between A, B and C, two observables that each person gives voice to can be taken together as the gradient of abstraction, i.e. the interface. The fact that in the example the three LoA’s happen to be non-overlapping does not need to be the

48_{Floridi (2011b), pp. 293-302; Floridi (2011a), especially chapter 3.} 49_{Brenner (2010)}

50_{Ibid., p. 112}

(17)

16 case. Different LoA’s could subsume the LoA of the three examples, resulting in perspective that contains more information. In that case we speak of a more concrete LoA, which is on a lower level that the others. Higher LoA’s are more abstract, in the sense that they leave out more of the system in their model than the more concrete, lower, LoA’s.

The point of using the method of levels of abstraction is that it works as ‘building’/’constructing’ a model of a system, from a specific perspective and for a specific purpose.52_{This coincides with}

Floridi’s further characterization of the method along the lines of the ontological commitment of a theory. A theory is constituted by the combination of a LoA, a model and the properties the model identifies with respect to the system under consideration. The theory commits itself ontologically by assuming a LoA on the one hand. On the other hand it is committed more specifically by the model. In other words, a LoA commits to certain types of objects with respect to a system, whereas the model actualizes the commitment to certain objects (Floridi calls them ‘tokens’) that correspond to the types in the LoA, see figure 1 below.

Figure 1

52_{Floridi (2011a), pp. 296-299}

(18)

17

2. Constructionism vs JTB

In the previous chapter an approach to the relation between knowledge and the world has been introduced, especially along the lines of Luciano Floridi’s view on situating constructionism in a maker’s knowledge tradition. As such, constructionism can be understood as a theory of knowledge, situated in between the ‘extreme’ positions of (commonsensical) realism on the one hand and the more sceptical constructivism on the other. This leads to the question of what kind of theory of knowledge constructionism is actually opposing. Floridi refers to a sort of mainstream tradition of naive realism in the contemporary scientific practice. He points to the user knowledge tradition, ascribing an ever-influencing role to it as if it is, perhaps unconsciously, the commonly presupposed way of looking at knowledge as a concept and the knowledge production process in practice. Therein, his ‘opponent’ remains rather abstract. However, in philosophy, the question of what knowledge is and how we should understand its production is a prominent one. That is to say, over the 20th century much work has been done to formulate a commonly accepted notion of knowledge. This notion is usually referred to as the Justified True Belief (JTB) account of knowledge.53

2.1 Knowledge as Justified True Belief

Mathias Steup, who has written an introductory work to contemporary epistemology54_{, and}

other textbook sources with respect to contemporary epistemology,55_{refers to JTB as the}

‘standard’ or ‘traditional account’ of analysis of knowledge. It is the commonly accepted starting point of thinking about knowledge. This ‘tripartite’ account of knowledge takes ‘truth’, ‘belief’ and ‘justification’ as necessary and sufficient conditions for a propositional knowledge56_claim.

The JTB account is defined as: S knows that p if and only if (i) p is true (ii) S beliefs that p, and (iii) S is justified in believing that p.57_{The first condition, truth, is a seemingly straightforward}

one. Obviously we do not wish to include false beliefs into our knowledge, if someone believes something that is false, it does not count as a knowledge claim. The second condition, belief, points towards the fact that if we do not believe a certain claim, we can hardly call it knowledge. Saying that the Earth is round while simultaneously claiming that I do not believe it to be so makes no sense. So, for S to know that p, S must believe that p. Thirdly, the justification-condition, deals with the potential for coincidentally believing that p is true. If S, by a mere stroke of luck, believes p and p turns out to be true, we do not wish to call it knowledge. In this sense, justification deals with the possibility of being correct by chance. In other words, justification deals with epistemic luck, by necessitating a rational justification of S’s belief that p is true. This constitutes the basic account of JTB. It should be noted that the brief description just given does not do justice to the complexity of all theorizing that has been done with respect to JTB. However, for the account of JTB to figure in this thesis, this confined version will be enough.

It is clear that the traditional account of knowledge, JTB, deals exclusively with propositional knowledge, S knows that p. In describing the position of ‘commonsensical realism’ I have related

53_{Steup (2017) SEP} 54_{Steup (1996)}

55_{see, for instance, Contemporary Debates in Epistemology (2014), ed. Steup, Turri & Sosa}

56_{Epistemology has largely focused on propositional knowledge. With Hans Radder in chapter 1.1.2 we} have seen that there are also epistemologists who emphasize other notions of knowledge, like

‘knowledge-how’. 57_{Steup (1996), p. 3}

(19)

18 it to the approach of understanding the different positions concerning knowledge along the ‘knowing-that’ and ‘knowing-how’ divide. Hans Radder explicitly employs this distinction to situate his view on knowledge as opposed to the traditionally accepted view on knowledge as justified true belief. Radder identifies seven issues with the JTB account58_{, I will not go into}

detail about each issue. However, overall his main issue with the JTB approach is that it focuses solely on that’. One the hand it excludes other types of knowledge, like ‘knowledge-how’, from the analysis of knowledge and on the other hand it excludes ‘know-how’ from the justification sphere. The latter issue is most illustrating in the sense that if S is to justify its belief of p, we are actually asking S to account for ‘how’ he knows that p. Thus, according to Radder, proponents of the JTB account of knowledge pay too little attention to know-how and are stuck in a tunnel of know-that, both on a general level and with respect to justification. That is why Radder argues that justification should be understood as a process that the epistemic agent is to perform.59

Floridi, as said, does not explicitly situate his conception of knowledge, as a concept most suitably understood through constructionism, opposing the JTB position. He refers to (naive) realism and the traditional view on knowledge in a more abstract sense. However, following Steup in pointing towards JTB as the ‘traditional’ view on knowledge, Floridi’s view can be situated in opposition to JTB. This raises the question of how Floridi deals with JTB and how his view on knowledge and its relation to reality square up with it. Furthermore, the question of why Floridi does not explicitly situate himself against JTB arises, since the JTB account is commonly accepted as the ‘traditional’ account of knowledge. In order to answer this question the remainder of this chapter introduces Floridi’s view on the position of JTB. Thereafter it discusses Floridi’s alternative to JTB as part of his philosophy of information. Finally, the role of constructionism is related to Floridi’s treatise of JTB and his alternative view. There I argue that there are points where the role of constructionism as a theoretical background issue is hard to deny.

2.2 How does Floridi deal with JTB?

As pointed out, Floridi does not explicitly situate his view of the maker’s knowledge approach as opposing JTB. Differently put, his notion of constructionism is not formulated as an alternative to JTB, but rather as opposing the mainstream tradition of viewing knowledge from a user’s perspective. In relating constructionism to JTB there are two possible courses of action in the context as just formulated. One the one hand JTB could be proven to be analogous to the ‘traditional view’ as Floridi identifies it. In order to do so, the JTB account as the ‘traditional’ account of knowledge should be analysed to the extent that its characteristics can be shown to correlate with the picture Floridi paints of the traditional approach to knowledge and a commonsensical realist view of the relation between knowledge and the world. Although this is not impossible, it is a laborious exercise and does not fall within the scope if this thesis as such. The other approach that can be taken is to accept the apparent relation between what Floridi understands as the commonly accepted view on knowledge and the JTB account. Accepting the possibility of taking JTB as an instance of the ‘traditional’ view that Floridi opposes with his maker’s perspective on knowledge allows for the question of how, if at all, Floridi deals with the JTB account of knowledge. It happens to be the case that Floridi does address the JTB account of

58_{Radder (2017)} 59_ibid.

(20)

19 knowledge and it plays an important role in the development of his argument. He does not oppose JTB along the user’s vs. maker’s knowledge tradition, but identifies it as an obstacle for his conception of information and of how, in his view, we move from information to knowledge. These strategies do not contradict each other per se. In light of the aim of this thesis, understanding PI in the context of the relation between knowledge and reality, I consider the lack of connection between Floridi’s view on JTB and his constructionist epistemology a loose end.

2.2.1 Gettierization of JTB

Although Floridi does not juxtapose his notion of constructionism against the JTB account of knowledge, he does pay attention to the great challenge that JTB faces, namely the Gettier-counterexample. In 1963, Edmund Gettier published his paper ‘Is Justified True Belief Knowledge?’60_{, where he argues that the three conditions of JTB are not sufficient in defining}

knowledge. There are many Gettier-examples and even a paper written which contains a recipe to construct your own.61_{In this thesis I will follow the work of Steup}62_{in laying out the basics of}

the Gettier-problem for JTB.

The Gettier-problem highlights two issues with taking JTB to account for propositional knowledge. Firstly, it is possible that S is justified in believing p, but in reality p happens to be false. Gettier illustrates this issue by proposing the following: suppose Smith is justified in believing that:

(i) Jones owns a Ford.

As Smith remembers (correctly) Jones has always owned a Ford and he has just had a ride to work, in a Ford driven by Jones. However, it happens that Jones has sold his Ford (the one he has owned as long as Smith remembers), and is now driving a Ford he has rented from the car-rental company.63_{Therefore, it turns out that proposition (i) is actually false. The second step}

that Gettier takes is applying the rules of deductive logic to proposition (i), he deduces the following three propositions:

(ii) Either Jones owns a Ford or Brown is in Amsterdam. (iii) Either Jones owns a Ford or Brown is in Utrecht. (iv) Either Jones owns a Ford or Brown is in Groningen.

Suppose that Brown is actually, by mere coincidences, currently in Amsterdam, without Smith being aware that he is. So Smith has no belief, nor justification that Brown is in Amsterdam. However, according to Gettier, Smith is justified in believing (ii). Note that, as far as Smith is concerned, he knows that (ii) for the same reason he thinks he knows that (iii) and (iv). Smith is not aware that there is a problem with (i), therefore he is correct in asserting (ii) due to sheer epistemic luck. Steup emphasizes that the ‘closure principle’64_{that Gettier employs has been}

60_{Gettier (1963)}

61_{Zagzebski (1994)} 62_{Steup (1996)}

63_{If renting a car transfers ownership we could also replace the rental example with a borrowing} example, for instance Jones has borrowed the car from his friend Johnson.

64_{This principle entails that justification from one proposition is transferred to another due to recognized} entailment, i.e. recognizing that proposition (i) entails the others based on deductive logic.

(21)

20 challenged, but that it remains plausible.65_{At the very least, this is theoretical plausibility,}

because, as Steup rightly points out, deducing one statement from another is a ‘truth-preserving’ process. In other words, deducing from a true statement cannot possibly provide a false statement, if that is the case, there is something wrong with the deduction. Ultimately, it is concluded that Smith is justified in having the true belief (ii), but we cannot speak of knowledge because it is a case of mere luck that (ii) is a justified true belief. The problem is illustrated by (iii) and (iv), of which Smith also claims to have a justified true belief, whilst they are not true at all. The view on the Gettier-example as it has just been presented is, similar to the treatise of JTB, a basic version of the theory which does not fully justify all the work that has been done on both subjects. However, as we move to Floridi’s approach to the ‘Gettier-problem’ and JTB, it will become clear that he identifies a fundamental problem in the latter of which Gettier’s counterexample is just a symptom.

2.2.2 Gettierization and the ‘coordinated attack’-problem

Floridi speaks of the tripartite analysis as providing necessary conditions, but not sufficient in defining propositional knowledge.66_{The conditions fail to define knowledge because they are}

not capable of dealing with ‘epistemic luck’. Epistemic luck in the context of JTB entails that an epistemic agent holds a certain belief in the form of a proposition, this proposition is true and the agent is justified in believing it to be true. However, due to circumstances unknown to the agent, the way in which he comes to believe the proposition is false. Therefore, the agent’s success in holding a piece of propositional knowledge is a case of mere luck. The argument is that having to depend on luck to hold some piece of knowledge, does not constitute as having knowledge or truly ‘knowing’ at all.

Floridi identifies the fundamental problem in JTB as shown by the Gettier-cases as a problem of ‘coordinated attack’. The coordinated attack problem is a part of epistemic logic also known as the ‘Two Generals’ Problem. It is a thought-experiment in computer science credited to Jim Gray67_{. Joseph Halpern}68_{sketches the thought-experiment as follows: suppose that there are}

two generals (A and B) and their armies, camped on two hills, flanking a valley where their enemy is situated. The generals do not have a standing plan of attack. It is clear to them that if they both attack at the same time, the enemy will be defeated. However, if they fail to coordinate their effort, the enemy will come out victorious. Therefore the generals attempt to communicate their plans through a messenger that takes some time to get from A to B, and vice versa.

Suppose that A sends a message stating ‘let’s attack at dawn tomorrow’. The messenger will have to carry this message to the other side, leaving him prone to interception by the enemy, or he might fail to deliver the message for other reasons. The point is that general A does not know if his message reaches B and even if B receives the message, sends a messenger back which reaches A, then B cannot be sure whether A received the confirmation. Logically speaking this leads to an impasse, since neither general attacks until the effort is coordinated with the other, and one of the generals is always in a state of not-knowing whether the plan is successfully coordinated. Halpern states that in a situation of ‘unbounded messaging’, the case that both A and B do not simultaneously know that the message is received; ‘common knowledge’ is impossible. In simpler terms, the fact that the possibility of failure to carry the message from A

65_{Steup (1996), p. 5} 66_{Floridi (2011a) p 210} 67_{Gray (1978)}

Constructionism and its role in philosophy of information: An analysis of the relation between knowledge and reality in Luciano Floridi’s philosophy of information