In 2005, Woods described the epistemic bubble as an immunized state of human cognition that compromises the awareness of the agent about her beliefs and knowledge. The idea of an immunized knower swung with the proposal advanced by Gabbay and Woods of constructing a practical logic and epistemology, which can actually define itself as agent-centered, goal-oriented, and resource-bound. In order to carry out this project, in this paper we will introduce a symmetrical view on the agent immunization, focused on the agent's missing awareness of her ignorance, also highlighting the importance of considering the actual agent as cogently ignorant, too. Proposition 3.2c The Error Abundance Thesis: Human beings make errors, lots of them.
In 2003 Gabbay and Woods officially proposed a program,
condensed in a series of volumes called “A Practical Logic of
Cognitive Systems”, to the aim of constructing new logical
models able to fill the gap between the logical and cognitive
representation of human agent and its “real” – eco-cognitive
multi-dimensioned – counterpart. The last volume of the
series, published in 2013, was aimed at drawing an empirically
“sensitive” and “aware” form of logic, able to deal with
actual reasoners’ cognitive performances
Proposition 3.2b The Cognitive Abundance Thesis: Human beings have knowledge, lots of it.
Proposition 3.2d The Enough Already Thesis. Human
beings are right enough about enough of the right things
enough of the time to survive and prosper
In this perspective the presentation of the “right enough
human being” is referred to the study of the cognitive
endowments of an “actual” agent, mostly focusing on her
knowledge and cognitive skills. The “ignorant” part is basically
described as an innocent tendency to commit errors (even if
lots of them) or treated in the light of fallacious reasoning.
We contend, following Proctor, that “ignorance is more than
a void”
On the contrary, the philosophical background that
informs the Naturalization of Logic resorts to Peircean
tradition: Peirce directly examined the problem of doubt and
tried to grasp its philosophical, epistemological, and
cogni2Certainly fallacies and heuristics have always been considered
the main door to step into the problem of “ignorance”, see, for
instance the classical
us, though very different ones. Belief does not make us
act at once, but puts us into such a condition that we shall
behave in some certain way, when the occasion arises.
Doubt has not the least such active effect, but stimulates
us to inquiry until it is destroyed
What Peirce defines as the irritation of doubt is an
unwanted state of mind caused by the loss of certainty in the
agent knowledge. We can consider this description as the
easier way to see the experience of a part of ignorance. The
individual desperately wants to escape from the condition of
doubt
In order to comprehend how Peirce’s epistemology grounds
the Naturalization of Logic we should leave for a moment the
analysis of doubt and briefly revisit the definition of belief.
Peirce describes the state of belief as having just three
properties: “first, it is something that we are aware of; second, it
appeases the irritation of doubt; and, third, it involves the
establishment in our nature of a rule of action, or, say for short
a habit”
The awareness of our belief state (that is, according to both Peirce and Woods, the only state that allows us “to know”) obviously is what makes us able to define ourselves “knowers”. In this perspective, the tendency to knowledge indicated by Woods in the proposition 3.2b is just a consequence of the awareness of how much we believe we know and how much we are able to learn. The second feature, the capacity of belief to appease “the irritation of doubt” is at the basis of the “Error Abundance” thesis, which composes the second item of Woods’ fallibilism. Believing is a pleasurable state, a state that calms the agent and gives her the cognitive resources to act. It is this practical advantage that makes it preferable to doubt, no matter if it is epistemically more convenient or less. The tendency to commit errors (and so of believing in an incorrect, or “fast and frugal”, statement instead of doubting it)4 of the actual agent is exactly derived from this unfortunate preference. Finally, the third condition of belief, which “involves the establishment in our nature of a rule of action, or, say for short a habit”, can be seen as the feature that seals the “Enough Already Thesis”. Even intuitively, believing to know something has two main consequences: a) it repels the irritation of doubt, making us sure about our own knowledge (sometimes compromising our ability to individuate errors in it) and, b) since belief gives us the possibility to act in the world upon a certain circumstance, we will be inclined to rely on the same belief as a principle for solving other similar circumstances. The “enough already thesis” does not affirm much more than the prevalence of the occurrence for our belief to be confirmed by a personal (more or less fortunate) experience.
The parallel between Woods’ Fallibilist principles and the Peircean definition of belief allows us, first of all, to confirm the knowledge-based perspective of Woods’ analysis itself. The actual agent is a knower, because she is also a believer. The fact she believes she knows implies the possibility of committing errors, but it does not compromise her epistemic status of knower. We contend, extending Woods’ characterization of the actual agent, that ignorance is more than a simple tendency to commit errors: taking advantage of the description of belief in Peirce’s work we can put down the equivalent three properties for doubt considering it as visible ignorance. Like belief, doubt is a state we are well aware of ; it is an unwanted and irritating state for the agent; it requires 3For clarity, we should mention that the neat difference between doubt and belief, in Peircean perspective, is grounded on and limited by the sensations and the practical awareness of the agent. It is not in question whether we can have unconscious beliefs or doubts but, according to Peirce, how and why the ones we perceive as affecting our behavior are epistemologically relevant.
4Here, thanks to Gigerenzer’s formula,
The most evident outcome is a definition of ignorance that
is already formulated in Woods’ theory. He defines
ignorance as “inferentially productive”, as a part of our cognition
that we can examine through fallacious but effective
inferential processes
The second and less evident consequence of introducing the doubt as “the visible part of our ignorance” in the fallibilist triad, is instead a sort of “negative” affirmation. Examining the epistemic status of ignorance at the conscious level, so addressing knowledge firstly as belief, and speaking of doubt as something that we do not recognize as belonging to our knowledge, we let the door open to the fact that an actual agent is not “simply” ignorant of what she is aware she doesn’t know. Ignorance is not completely equivalent to doubt, it is not just a missing piece of our cognition, something that the agent knows she does not know. Doubt can be perceived as the frame of our ignorance, but it is the simple consequence of our fallible cognition. Of course the project toward the Naturalization of Logic has already proposed to extend the field in which knowledge can be investigated, contending the importance of the examination of the errors of reasoning and their “positive” aspects. Something more can be said: since ignorance corresponds to something that goes, at least partially, beyond the cognitive sight of the agent, we should investigate it as the Naturalization of Logic aims at explaining the limit of our knowledge: far beyond the selfperspective of the actual agent.
The “negative” affirmation permits us to have an “Enough
Already Thesis”, less indulgent with respect to the agent’s
actual status. As we have already said, the psychological and
emotional component of doubt makes its experience
repulsive for the agent. So, if even the visible part of ignorance is
hard to be managed by the agent, the part of ignorance that
falls beyond her control (or her will) must be extremely
difficult to reach. The “Enough Already Thesis” displays the
capacity of human beings to be right enough about enough
of the right things enough of the time to survive and prosper
Hence, instead of focusing on the confidence in the
“Enough Already Thesis” (that we can condense in “we are
able to survive, after all”), all the added caveats demand a
deeper questioning on the tendency of human beings to avoid
a complete awareness of their own ignorance. It is not
unreasonable asking how and why this is functional, for instance.
The examination of these important issues is already
displayed in what Woods called the Epistemic Bubble
The idea of the Epistemic Bubble originates from the analysis of both purpose and ending of the state of doubt, albeit it remains focused on the analysis of the state of belief. It is based on the assumption that the state of belief is not just pleasant, but also fallible and uncertain. In its essence, Woods’ Bubble Thesis focuses on the relation between the complex of beliefs an agent has and her awareness as regard as either their correctness or unsteadiness. This suggests that the agent’s mechanism of belief formation can provide an easy way out to the Peircean irritation of doubt through a systematic ascription of knowledge concerning a mere belief, immunizing the agent from being able to spot the difference (so letting the agent think she knows something when she merely believes she does). In order to utterly understand the potentialities of this idea, we have to introduce two dichotomies that Woods indicates as substantial. The first between what we could consider a broad definition of knowledge related to the Peircean state of belief. The second concerning the important distinction between the first and third-person perspective of the agent.
As already mentioned, belief in Woods’ theory
corresponds to the Peircean definition: it is the sole state that solves
the irritation of doubt and brings peace to the cognitive
unsteadiness of the agent; fundamentally, it is a state that calms
the agent’s mind. Knowledge, instead, is defined as a “kind of
case-making. One knows that P only if one has one’s disposal
a case of requisite strength to make for P”
The entanglement between knowledge and belief drives our argumentation to the difference between the first and the third-person perspective. Indeed, for the agent is kind of easy to say if someone else knows or thinks she knows something. That is to say, from the third-person perspective one can tell the difference between a belief that stands for an actual knowledge attainment and a belief that just brings about some cognitive relief to an irritating state of doubt. The agent can judge if someone else’s is either effective knowledge or mere confidence. From the first-person perspective, the difference is instead blurred, due to the fact a belief state entails the occurrence of knowledge. This is an entanglement indeed recognized as the focus for the asymmetry between first and third-person perspective. Whenever the agent knows something, she is compelled to believe she knows it. But, since the attainment of knowledge is different from the establishment of a belief, she can believe she knows something even when she does not. This distinction between knowledge and its mere ascription is visible only in a third-person perspective.
Hence, while in the first-person perspective a reliable belief is always claimed as knowledge, in the third-person perspective the proposition can be judged as potentially verified or erroneous. Thus, in the case of the first-person perspective there is not a clear distinction between knowing and believing in something, even if it is pretty clear in the case of the agent’s third-person perspective. At the same time, in the case of the first-person perspective the state of belief represents not only the way the agent can experience some relief from the irritation of doubt, but also the unique possibility for the agent of attaining any sort of knowledge. This idea is better expressed in Woods terms in the Proposition 6: (The Downside of Belief). Belief is both a condition of knowledge and an impediment to its attainment.
So, albeit the fact that there is a solid difference between
the epistemological status of belief and knowledge, the agent
cannot be aware of this distinction when she has to deal with
her own cognition. Hence, in its essence, the epistemic
bubble is configured as a first-person knowledge-ascription,
performed by the knowing agent, to whom the difference
between knowing something and thinking she knows that same
thing is unapparent – and the tension that may arise is always
solved in favor of the former
Woods describes the epistemic bubble as an autoimmune mechanism of the agent. The naiveness of the agent about her own cognition is directed by the same system that at the same time permits her to attain any type of knowledge. Belief, as a cognitive structure, is in primis a tool that gives her the possibility of taking action into the world. If the agent could not be sure about what she thinks she knows, she could not take any decision and she would be constantly in a state of doubt and struggle. The autoimmune mechanism helps her out from the freezing state of doubt but does not provide a safe exit from it.
As we can see, the epistemic bubble as an autoimmune mechanism concerns the limits of the attainment of knowledge, its entanglement with the state of belief, and the unapparent distinction between the two in the first-person perspective. As we have said in the case of the original Fallibilist principles mentioned in the previous section, the idea of epistemic bubble is profoundly connected with the definition of belief offered by Peirce. Using a similar connection, in order to shift the focus on the limits of ignorance-recognition, we must reconsider Peirce’s doubt in the light of the autoimmune mechanism described above.
Given the fact belief and knowledge are connected in the firstperson perspective, but way far from each other in the thirdperson view, we can formulate the same consideration in the case of doubt and ignorance. The “negative” affirmation in the Fallibilist principles is oriented to highlight the distinction between doubt and ignorance, but this separation is manifest only in a third-person perspective. In the third-person perspective, doubt presents the character of being a state of irritation for the subject, a push for inferential reasoning, and, mainly, a state she is aware of : it is a frame of the ignorance of the subject in those limits. The proper ignorance of the agent is beyond the frame of her doubts. It is something the agent cannot consider in first-person perspective. At the same time, the only “visible part of ignorance” for the first-person perspective can entirely frame the ignorance of the agent.
This relation is clear when we think about the possibility of describing how we ignore something. The only method that we can apply is to frame our ignorance, speaking about the propositions we doubt to be true, the situations we are not certain about, and the collection of data we are not sure if they are reliable or not. But these data are just what we consider part of our ignorance. They cannot be even close to the propositions we are not informed of, the situations out of our sight, and the collection of data we are not aware of. These data are part of our ignorance, but we cannot reach them through our doubts. At the same time, doubt is the only cognitive tool that permits us to grasp pieces of ignorance and let us admit that there is something out of our reach.
So, exactly as in the epistemic bubble, albeit the fact that there is a solid difference between the epistemological statuses of doubt and ignorance, the agent cannot be aware of this distinction when she has to deal with her own cognition. Consequently, we can describe the ignorance-based bubble as a missing-ascription of ignorance, performed by the agent, to whom the difference between ignoring something and doubting is unapparent.
This structure is also an autoimmune mechanism of the agent. Doubt, the only tool that permits the agent to investigate a part of her ignorance, makes also impossible for the agent to distinguish the amount of actual ignorance she possesses from what she is just able to recognize. At the same time, without this autoimmune system we would never leave the state of doubt. The ignorance-based autoimmune mechanism illustrates the ignorance about one’s own ignorance as the only possible condition for the attainment of any kind of knowledge in more or less uneasy condition. The immunization to ignorance is an indefeasible mechanism of human cognition as well it is the epistemic bubble. They simply define the borders of possibility for first-person perspective agents to modify their own epistemological status.
By considering the cognitive state of doubt, we can extend our analysis also considering Woods’ thesis about truth. As we will better illustrate, the analysis of the epistemic bubble leads to the affirmation that the truth, for the first-person perspective, is a fugitive property. In brief, the difficulty for the agent to distinguish the difference between what she knows and what she believes, impairs her possibility to reach and recognize truth. Using the same association in the case of the analysis of the agent’s immunity to her own ignorance we can arrive to a similar consideration regarding her capacity to reach and recognize the entireness of ignorance beyond the frame of her doubts.
The autoimmune system of the epistemic bubble makes the
attainment of truth a relatively impossible task from the
firstperson perspective, adding a veil of skepticism to the
cognitive analysis. This is clearly stated in Proposition 15, stating
the Fugitivity of Truth: “Within epistemic bubbles, truth is
a fugitive property. That is, one can never attain it without
thinking that one has done so; but thinking that one has
attained it is not attaining it”
As we have already mentioned, the ignorance that the agent can perceive is just defined through her doubts, and her doubts can picture just a small portion of her ignorance. The disparity between the two parts of her ignorance can be illustrated using the Freudian metaphor of the iceberg: the portion apparent to the subject is just a small piece with respect to the whole structure. For this reason the missing-ascription of ignorance plays a role analogous to that of the epistemic bubble in the mechanism of creation and revision of beliefs. It assures a cognitive status of certainty about the agent ignorance that permits the agent to be confident in her choices and knowledge. The agent, not being able to see how much she ignores, considers the attainment of answers concerning her doubts a concrete way to remove her ignorance piece by piece.
The role of confidence is part of the autoimmune mechanism as much as the proper ignorance embubblement. The embubblement allows the agent to consider what is part of her doubts as the entire amount of her ignorance and her purpose will be to remove it as much as possible. In this sense, the role of the missing-ascription of ignorance is fully motivational. But the more effective consequence in the agent’s cognition is the self-representation that the agent constructs in first-person perspective: indeed, there is a tendency to consider the knowing or ignorant self as a controllable part. The agent is fully aware of both the state of belief and of doubt, which are the only vehicles for her attainment of propositional/sentential knowledge and her partial awareness of ignorance. In the following subsection we will show that these partial recognitions drive the agent to formulate a sort of Homunculus Fallacy, when she tries to depict her epistemological state.
In the case of the first-person perspective we have illustrated above the epistemic bubble provides two main illusions. The first illusion is strictly related to the epistemic dimension of the bubble: it provides the belief-based ascription of knowledge even when that knowledge is not entirely attained. The second illusion is related to the cognitive and emotional outcome of the bubble: it makes the agent convinced of being aware of the knowledge she possesses, even when she should not.
The same deceptive double effect also emerges from the missing-ascription of ignorance. On the one hand, it provides the agent the conviction that she is ignoring just a specific sort of data, categorizable in the framework of her first-person perspective. On the other hand it gives the agent the illusion of being able to have a clear view of her own ignorance. In both cases the agent is naively assured about her cognition. She thinks herself able to see her knowledge and her ignorance as they were, respectively, sets of attained or missing propositions. The agent is deluded into being, absurdly, in an indifferent position about her own ignorance/knowledge structure. This effect can be pictured as a sort of Homunculus Fallacy. The subject thinks herself almost as a double being: one part of her knows and ignores and another part can spot how much she knows and how much she ignores.
The fairly hidden Homunculus Fallacy is clear: the
autoimmune mechanism suggests that the agent can judge about the
attainment of knowledge or the perception of ignorance, as
if the judgment belonged to a distinct part, which directly
knows or ignores. This illusory distinction allows us to also
consider the property of just apparent corrigibility of the
bubbles: “Since each of us is in his own epistemic bubble, the
distinction between merely apparent correction and genuinely
successful correction exceeds the agent’s command. [. . . ]
Within an epistemic bubble the distinction between
beliefchange and belief-correction is also “resolved” in favor of
the latter”
As it can be imagined, a similar structure is present in the account of ignorance-based bubbles: the “end” of a missingascription of ignorance happens when the obtained answer to a given doubt is just apparent. The missing-ascription of ignorance shifts to another problem, which arises in the presence of new collected information. While for the epistemic bubble there is a distinction between belief-change and beliefcorrection, which is resolved in favor of the latter, in the case of the ignorance-based bubble there is a distinction between change of doubt and ignorance-removal that is resolved in favor of the second. In conclusion, the autoimmune system provides the agent with an efficient and improvable mechanism of belief and doubt change without the loss of confidence in self-awareness.
The introduction of the problem of ignorance in the
framework of a naturalization of logic involves problematic issues
regarding the epistemological status of the “real agent”. First
of all we have added to Woods’ Fallibilist principles what
we have called the negative affirmation: this move rendered
possible the examination of the naiveness of the individual
agent about her own cognition, shifting the attention to the
state of doubt (defined by Peircean dynamic) instead of belief.
Thanks to this change of perspective, a new subtle
reinterpretation of Woods’ “epistemic bubble” has favored the
elicitation of that autoimmune mechanism that affects not only the
system of belief creation and revision of a human agent –
considered not able to distinguish what she knows and what
she only thinks she knows – but also the relationship between
doubt and ignorance-recognition. As belief is “the condition
of knowledge and the impediment of its attainment”