Information Evaluation

Philippe Capet, Thomas Delavallade

• Publisher: Wiley-ISTE
• Year: 2014
• ISBN:
  9781118898970
• Edition: 1
• Topic: Computer Science
• Subtopic: Information Technology
• Index: Computer Science

During the reception of a piece of information, we are never passive. Depending on its origin and content, from our personal beliefs and convictions, we bestow upon this piece of information, spontaneously or after reflection, a certain amount of confidence. Too much confidence shows a degree of naivety, whereas an absolute lack of it condemns us as being paranoid. These two attitudes are symmetrically detrimental, not only to the proper perception of this information but also to its use. Beyond these two extremes, each person generally adopts an intermediate position when faced with the reception of information, depending on its provenance and credibility. We still need to understand and explain how these judgements are conceived, in what context and to what end.
Spanning the approaches offered by philosophy, military intelligence, algorithmics and information science, this book presents the concepts of information and the confidence placed in it, the methods that militaries, the first to be aware of the need, have or should have adopted, tools to help them, and the prospects that they have opened up. Beyond the military context, the book reveals ways to evaluate information for the good of other fields such as economic intelligence, and, more globally, the informational monitoring by governments and businesses.

Contents

1. Information: Philosophical Analysis and Strategic Applications, Mouhamadou El Hady Ba and Philippe Capet.
2. Epistemic Trust, Gloria Origgi.
3. The Fundamentals of Intelligence, Philippe Lemercier.
4. Information Evaluation in the Military Domain: Doctrines, Practices and Shortcomings, Philippe Capet and Adrien Revault d'Allonnes.
5. Multidimensional Approach to Reliability Evaluation of Information Sources, Frédéric Pichon, Christophe Labreuche, Bertrand Duqueroie and Thomas Delavallade.
6. Uncertainty of an Event and its Markers in Natural Language Processing,
Mouhamadou El Hady Ba, Stéphanie Brizard, Tanneguy Dulong and Bénédicte Goujon.
7. Quantitative Information Evaluation: Modeling and Experimental Evaluation,
Marie-Jeanne Lesot, Frédéric Pichon and Thomas Delavallade.
8. When Reported Information Is Second Hand, Laurence Cholvy.
9. An Architecture for the Evolution of Trust: Definition and Impact of the Necessary Dimensions of Opinion Making, Adrien Revault d'Allonnes.

About the Authors

Philippe Capet is a project manager and research engineer at Ektimo, working mainly on information management and control in military contexts.
Thomas Delavallade is an advanced studies engineer at Thales Communications & Security, working on social media mining in the context of crisis management, cybersecurity and the fight against cybercrime.

Chapter 1

Information: Philosophical Analysis and Strategic Applications

1.1. Introduction

Nowadays, it is commonly considered that we live in an information society. Our civilization is abundantly fed by information and communication technology (ICT). In the military context, information has always been an invaluable and sought-after commodity to which special services are devoted. Leaders in all sorts of contexts – particularly military – have, for centuries, been aware of the importance of information. It is all the more paradoxical that in the area of philosophy, thinkers have largely ignored the concept of information. It appears only incidentally in philosophers’ writings. It was not until thermodynamics, cybernetics and mathematical theorization of information took off that philosophers finally became interested in it, and even then, it was difficult to find a structured theorization centered on the notion of information.

In this chapter, we shall focus primarily on two aspects of this topic. To begin with, we shall look at the definition of an operational concept of information. In order to do so, following a brief state of the art on current thinking about the notion of information, we turn our attention more specifically to the genealogy of the term, before going on to discuss two of the three thinkers whom we believe are the most important in the area of philosophy of information: Paul Grice and Fred Dretske. The work of these two philosophers will help to sculpt the precise definition of what information is, which will then be used in the second part of this chapter.

In this second part, we shall focus on the military domain, and see how the concept of information is used in that domain. In doing so, we shall present and critique the ideas of the third great contemporary thinker on information: Luciano Floridi.

Let us make it clear right now that while these philosophers help us to refine the concept of information which we shall use throughout this book, our ideas differ from theirs on a crucial point, because we reject the alethic conception of information – i.e. the idea that information must necessarily be true.

1.2. State of the art in philosophy

It is noteworthy that in the esteemed French reference dictionary of philosophical vocabulary – the Lalande – the term “information” does not appear. The work includes only the entry “to inform”, which is given in the scholastic sense – that of “giving shape to, organizing matter”. This view of the term comes directly from Plato’s Timaeus, where the demiurge imposes order on a disorganized physical world by giving it a form reflecting the order reigning over the world of Ideas. Apart from this primary meaning of the verb “to inform”, the Lalande recognizes a derived meaning, which is: “making somebody aware of something”. Granted, the Lalande dates from 1927, but none of the later editions, including that from 2006, have the entry “information” – at least not in the body of the text. It is only half a century after the Lalande first appeared that the tenth edition included a separate supplement containing the entry “information”. It is given a meaning derived from cybernetics: “an element of knowledge conveyed by a message which is its vehicle and of which it constitutes the meaning” [LAL 06]. In this definition, we can see the hallmarks of the work of Shannon. The persistent practice of affording only a very limited place to the concept of information in a philosophical dictionary which is highly influential – at least in the French-speaking world – highlights the recent and underdeveloped nature of the philosophy of information. It should be noted that this neglect of the concept of information is not specific to the French-speaking world. If we look at Simon Blackburn’s Oxford Dictionary of Philosophy [BLA 08], we discover that even the second edition, revised in 2008, does not contain an entry for Information. It only has an entry for Information theory. This gives an indication of why philosophers gradually lost their indifference toward the concept of information during the latter half of the 20^th Century. The domain known as “philosophy of information” first developed thanks to cybernetic research and the Mathematical Theory of Communication (MTC), then moved forward in the wake of philosophical explorations of the concepts of meaning and knowledge, and finally flourished due to the current development of the so-called “information society”.

1.2.1. History

The Lalande shows that, historically, the word “information” arose later than the verb “to inform”, from which it is derived.¹ In the work of both Plato and Aristotle, there is the idea that information is necessary for the passage of the materia prima from pure potentiality to actuality. This would remain the prevailing definition from Ancient times until the 17^th and 18^th Centuries, with the British empiricists (John Locke, David Hume). It is thanks to these empiricists that the verb “to inform” lost its original meaning and came to be understood in its current sense. The empiricists, abandoning the rationalistic credence whereby our minds come into this world already holding some innate ideas, attempted to explain how these ideas come into being. If the mind is not informed in advance, i.e. if it is not molded into a form, by the demiurge who leaves his mark upon it, imposing his seal which is constituted by innate ideas [DES 92], then we need a theory about the origin of our ideas. The choice of the empiricists, led by Locke [LOC 90] was to consider the mind as a blank slate upon which the world itself inscribes ideas when we interact with it. The second sense of the verb “to inform” and the notion of information as it is understood nowadays came from these considerations. The human mind is not naturally informed by any sort of demiurge; it receives its form from the outside world, which leaves its mark upon it. When we interact with the external world, it transmits to our mind some knowledge about reality. That is to say that the world informs our mind by imbuing it with certain ideas, certain knowledge. This gives us the derived meaning of the verb “to inform”, which no longer means simply “to model” or “to mold”, but also “to convey knowledge”. Information, therefore, is no longer merely the act consisting of shaping a material object, but also the thing that is conveyed during the production of ideas. This second sense of the word “information” is the prevailing one now, and is the one of interest to us here.²

In spite of the introduction of this second sense of the word “information” by the empiricists, it should be noted that no “philosophy of information” worthy of that title would develop at that time. Information was to remain a blind spot for philosophy and epistemology until the latter half of the 20^th Century. In the 19^th Century, there would be a tacit revival of thinking about information, with the work done on thermodynamics, and in particular the link, demonstrated by Boltzmann in 1894, between entropy determined by the second law of thermodynamics and the amount of information accessible. In the 1950s, Claude Shannon used Boltzmann’s work as the basis for the first mathematically rigorous definition of the concept of information. Paradoxically, it was only in the wake of Shannon’s work that philosophers began to turn their attention to the notion of information. Of these philosophical works, particular mention ought to be given to the distinction drawn by Paul Grice between natural meaning and non-natural meaning, the use of the notion of information by Dretske in support of an externalist epistemology, and the recent work of Luciano Floridi, who has truly revivified the field of philosophy of information (PI).

1.2.2. Information at the crossroads between epistemology and philosophy of language

There are two main philosophical domains wherein the notion of information plays a crucial role: philosophy of language and naturalized epistemology. In philosophy of language, an important question is how to define the meaning of the expressions of the language. How can we say that someone understands a word or an expression? Shall we rely on behavioral indicators? Is to understand a sentence to be able to translate it into another language? For technical reasons neither of these solutions were satisfactory. This problem was solved by Donald Davidson³, who posited that to understand an expression is to be able to state its truth conditions: this is the truth-conditional conception of semantics. In Davidson’s view, the purpose of language is to provide us with information about the state of the world. Knowing the truth conditions of an expression is to know what state of the world would render that expression true. We can therefore consider that all expressions inform us about the state of the world. Hence, comprehension is nothing other than reception of the information encoded in language. One might therefore have expected logicians and philosophers of language who accepted a truth-conditional view of semantics to develop a school of thought centered on the notion of information. It did not happen. It was not until Dretske⁴ reflected upon Paul Grice’s work (see [GRI 89]) that information came in, one might say through a hidden door, to philosophy of language and epistemology.

1.2.2.1. Meaning and information in Grice’s work

It is precisely because Grice did not blindly accept the conventional view held by analytical philosophers – according to which “the essential business of language is to assert or deny facts” [RUS 22] – that he reintroduced the notion of information to philosophy of language, albeit in a roundabout way. In a 1957 article, he revisited the notion of meaning. As an ordinary language philosopher, he refuted the thesis according to which the only legitimate use of language was to describe states of the world that can be evaluated. He distinguishes two notions of meaning, of which he says the following two utterances are paradigmatic:

1) Those spots mean measles;

2) Those three rings mean that the bus is full.

Grice shows that the notions of meaning at play in the first and the second example are very different. If we consider the first example, the link between the spots and the measles infection is automatic and indisputable. For instance, it would be nonsensical to say “those spots mean measles but there is no measles infection”. If the eruptions of the skin are indeed indicative of measles, their link with the disease is such that it is impossible to have them without being afflicted with measles. Furthermore, as Grice notes, it is impossible to deduce from “those spots mean measles” the consequence “measles is meant by those spots”. Note that such a deduction would be entirely possible in the example of the bus, where we can understand sentence (2) as being equivalent to “the driver means, by those three rings, that the bus is full and about to depart”.

Grice refers to the first sense of the verb to mean as natural meaning. In natural meaning, there is an unbreakable link between the two things being connected. One of these terms is an indicator of the other. The presence of a certain type of spots necessarily indicates that the person bearing them is infected with measles. In the case of natural meaning, it is (so to speak) intrinsically that the subject indicates the occurrence of what is meant. Hence, it would be contradictory to affirm at the same time that A means B and that A is occurring without B being the case. This is what led Dretske, commenting on this notion, to say that “natural meaning is information” [DRE 08]. Because natural meaning expresses an inalienable linking of two different things, it is information. This enables us to establish an initial link between the notions of meaning and information. One might, at first glance, imagine that meaning and information are one and the same thing, as suggested by considering only natural meaning. If we take account of the indicative link which exists within natural meaning, we can test whether the connection characteristic of the notion of natural meaning is always true. If, for instance, I say that the number of rings in a tree trunk indicates the age of the tree, I am saying that those rings give us information about the age of that tree, and that the age calculated using this method is always correct. The number of rings therefore literally means the age of the tree. Now let us look at the second sense of meaning envisaged by Grice to clarify the notion of information.

We saw earlier that in the case of a ringing tone on a bus, it is possible to deduce from the sentence “Those three rings mean that the bus is about to depart” the sentence “Somebody is indicating, by those three rings, that the bus is about to depart”. Therein lies the main difference between natural meaning, which associates a symptom with its cause, thus giving us objective information, and the second sort of meaning Grice examines. In this second sort of meaning, the correlated variation between the two entities being linked is mediated by the human mind. There is necessarily a person who is using one entity to indicate the other. If, for instance, the public transport company instructs its drivers to ring three times to signal that the bus is full, that triple ring will come to mean that the bus is full. Each time a person who is aware of the local mores hears those three rings, he will inevitably understand that the bus is full. However, as Grice stresses, in this case there is not necessarily any causal link between the fact that the bus is full and the three rings. For example, there is no contradiction in exclaiming that “Those three rings mean that the bus is full, but the bus is only half full!” In the case of natural meaning, there is an impossibility of having A without having B given that A means B. For example, it is impossible to have a tree whose trunk has sixteen rings but whose age is not 16 years. On the other hand, the bus driver can ring three times whether or not the bus is full. This demonstrates that this second sense of meaning has an element of the arbitrary. It is for this reason that Grice dubs this non-natural meaning. With non-natural meaning, a connection is made by the speaker between two entities that otherwise would not necessarily have varied correlatively. It is the very fact that the person using A to mean B, non-naturally, associates the two so that A ends up meaning B.

What are the exact characteristics of non-natural meaning? One possible definition would be that non-natural meaning is any association made between items that are not necessarily covariant, and intended to convey information. For instance, said Grice in this case, we would speak of non-natural meaning when a perpetrator leaves clues incriminating somebody else at the scene of his own crime. After all, the aim of such a tactic is to convey information (the culpability of an individual) by combining two items (in this case the crime committed and the clues left behind). Returning to the fictitious characters embroiled in the scenario around which this book is woven, suppose that General San-Gerio leaves the fingerprints of his enemy, General Quieto, on the weapon he has used to commit a murder. Would we, in this case, say “The presence of his fingerprints means that General Quieto is guilty”? Not really, Grice would maintain. We could certainly say “General San-Gerio is attempting to signify by these fingerprints that General Quieto is guilty”, or more accurately, “General San-Gerio is attempting, by leaving these fingerprints, to convey false information incriminating General Quieto”. Yet it would be a fallacy to affirm that: “The presence of his fingerprints means that Quieto is guilty”. Nevertheless, San-Gerio undeniably has the intention of conveying information indicating that Quieto is guilty. What this demonstrates, according to Grice, is that the intention of the person who makes the two signals covary is not sufficient to bring out non-natural meaning. What, then, do we need in order to be able to speak of non-natural meaning? It is certainly essential, in the case of non-natural meaning, that someone attempt to communicate information. Yet this is not sufficient. We still need to take account of the fact that recognition of that intention is essential to the accomplishment of the transmission of information. If we look again at the case of the ringing bell on the bus, the three rings come to mean that the bus is full because, not only is the driver instructed to trigger those rings to indicate to the customers that the bus is full, but also because everybody knows that the sound is triggered to that effect. Hence, there is a coordination of the mental states (here, beliefs and intentions) of the different members of the relevant community, which results in a covariation between the sounds of the bell and the filling of the bus. Given that the non-natural lexical meaning emerges because of a coordination of the mental states of the people involved, it is not a case of transmission of an objective and necessarily true piece of information. In a given system, we can have an item A which c...

Table of contents

• Cover Page
• Table of Contents
• Title Page
• Copyright
• Foreword
• Introduction
• Chapter 1: Information: Philosophical Analysis and Strategic Applications
• Chapter 2: Epistemic Trust
• Chapter 3: The Fundamentals of Intelligence
• Chapter 4: Information Evaluation in the Military Domain: Doctrines, Practices and Shortcomings
• Chapter 5: Multidimensional Approach to Reliability Evaluation of Information Sources
• Chapter 6: Uncertainty of an Event and its Markers in Natural Language Processing
• Chapter 7: Quantitative Information Evaluation: Modeling and Experimental Evaluation
• Chapter 8: When Reported Information is Second Hand
• Chapter 9: An Architecture for the Evolution of Trust: Definition and Impact of the Necessary Dimensions of Opinion Making
• List of Authors
• Index