Many philosophers have stumbled across the possibility of the BIV scenario — that is, that our experience of reality could be a result of our brains being in a vat, connected to electrodes which are controlled by some mad scientist, thus creating the illusion of a physical body, universe, etc. Many variations of this idea have been presented which appear to be similar to or based off of the “Evil Demon” proposed by Descartes in his Meditations on First Philosophy back in 1641.
There is no way to prove or disprove this notion because all of our “knowledge” (if any exists at all depending on how we define it) would be limited to the mad scientist’s program of sensory inputs, etc. What exactly is “knowledge” ? If we can decide on what qualifies as “knowledge”, then another question I’ve wondered about is:
How do we modify or shape the definition of “knowledge” to make it compatible in the BIV scenario?
Many might say that we have to have an agreed upon definition of “knowledge” in the first place before we modify it for some crazy hypothetical scenario. I understand that there is disagreement in the philosophical/epistemological community regarding how we define “knowledge”. To simplify matters, let’s start with Plato’s well known (albeit controversial) definition which basically says:
Knowledge = Justified True Belief
It would help if Plato had spelled out an agreed upon convention for what is considered “justified”, “true”, and “belief”. It seems reasonable that we can assume that “belief” is any proposition or premise an individual holds to be true. We can also assume that “true” implies that the belief is factual — that is, that the “belief” has been confirmed to correspond with “reality”. The term “justified” is a bit trickier and concepts such as “externalism vs. internalism”, probability, evidentialism, etc., indeed arise when theories of justification are discussed. It is the “true” and “justified” concepts that I think are of primary importance (as I think most people if not everyone can agree on what a “belief” is) — and how they relate to the BIV scenario (or any similar variation of this scenario) in our quest for properly defining “knowledge”.
Is this definition creation/modification even something that can be accomplished? Or will any modification necessary to make it compatible in the BIV scenario completely strip the word “knowledge” of all practical meaning? If the modification does strip the word of most of its meaning, then does that say anything profound about what “knowledge” really is? I think that, if anything, it may allow us to step back and look at the concept of “knowledge” in a different way. My contention is that it will demonstrate that knowledge is really just another subjective concept which we can claim to possess based on nothing more than an understanding of and adherence to definitions, and the definitions are based on hierarchical patterns observed in nature. This is in contrast to the current assumption that knowledge is some form of objective truth that we strive to obtain with a deeper meaning that may in fact (but not necessarily) be metaphysical in some way.
How does the concept of “true” differ in the “real world” vs. the BIV’s “virtual world”?
Can any beliefs acquired within the BIV scenario be considered “true”? If not, then we would have to consider the possibility that nothing “true” exists. This isn’t necessarily the end of the world, although for most people it may be a pill that is a bit too hard to swallow. Likewise, if we consider the possibility that at least one belief in our reality is true (I’m betting that most people believe this to be the case), then we must accept that “true beliefs” may exist within the BIV scenario. The bottom line is that we can never know if we are BIVs or not. So this means that if we believe that true beliefs exist, we should modify the definition to apply within the BIV scenario. Perhaps we can say that something is “true” only insofar as it corresponds to “our reality”. If there are multiple levels of reality (e.g. our reality as well as that of the mad scientist who is simulating our reality), then perhaps there are multiple levels of truth. If this is the case then either some levels of truth are more or less “true” than others, or they all contain an equally valid truth. This latter possibility seems most reasonable as the former seems to negate the strict meaning of “true”, since it seems that something is either true, false, or neither (if a third option exists). There is no state that lies in between that of “true” and “false”. By this rationale, our concept of what is true and what isn’t is as valid as the mad scientists. Perhaps the mad scientist would disagree with this, but based on how we define true and false, our claim of “equally valid truth” seems to be reasonable. This may illustrate (to some) that our idea of “true beliefs” does not hold any fundamentally significant value as they are only “true beliefs” based on our definitions of particular words and concepts (again, ultimately based on hierarchical patterns observed in nature). If I define something to be “true”, then it is true. If I point to what looks like a tree (even if it’s a hologram and I don’t know this) and say “I’m defining that to be a “tree”, then it is in fact a tree and I can refer to it as such. Likewise, if I define it’s location to be called “there”, then it is in fact there. Based on these definitions (among others) I think it would be a completely true statement (and thus a true belief) to say “I KNOW that there is a tree over there”. However if I failed to define (in detail) what a “tree” was, or if my definition of “tree” somehow excluded “a hologram of a tree” (because the definition was too specific), then I could easily make errors in terms of what I inferred to know or not know. This may seem obvious, but I believe it is extremely important to establish these foundations before making any claims about knowledge.
Turning our attention to the concept of “justified”, one could argue that in this case I wouldn’t need any justification for the aforementioned belief to be true as my belief is true by definition (i.e. there isn’t anything that could make this particular belief false other than changing the definitions of “there”, “tree”, etc.). Or we could say that the justification is the fact that I’ve defined the terms such that I am correct. In my opinion this would imply that I would have 100% justification (for this belief) which I don’t think can be accomplished in any way other than through this definitive method. In every other case of justification we will be less than 100% justified, and there will only be an arbitrary way of grading or quantifying that justification (100% is not arbitrary at all, but to say that we are “50% justified” or “23% justified” or any other value other than that 100% would be arbitrary and thus meaningless). However, I will add that it is possible to have a belief that is MORE or LESS justified than another, but we could never quantify how much MORE or LESS justified it may be, nor will we ALWAYS be able to say that one belief is more or less justified than another. Perhaps most importantly regarding the importance of “justification” with regard to knowledge, we must realize that our most common means of justification is through empiricism, the scientific method, etc. That is, our justification seems to be based inductively on the hierarchical patterns which we observe in nature. If certain patterns are observed and are repeatable, then the beliefs ascertained from them will not only be pragmatically useful, but they will be most justified.
When we undergo the task of defining knowledge or more specifically the task of analyzing Plato’s definition, I believe that it is useful for us to play the devil’s advocate and recognize that scenarios can be devised whereby the requirements of “justified”, “true”, and “belief” have been met and yet no “true knowledge” exists. These hypothetical scenarios can be represented by the well known “Gettier Problem”. Without going into too much detail, let’s examine a possible scenario to see where the problem arises.
Let’s imagine that I am in my living room and I see my wife sitting on the couch, but it just so happens that what I’m looking at is actually nothing more than a perfect hologram of my wife. However my wife IS actually in the room (she’s hiding behind the couch even though I can’t see her hiding). Now if I were to say “My wife is in the living room”, technically my belief would be “justified” to some degree (I do see her on the couch after all and I don’t know that what I’m looking at is actually a hologram). My belief would also be “true” because she is actually in the living room (she’s hiding behind the couch). This could be inferred as fulfilling Plato’s requirements for knowledge: truth and justification. However most people would argue that I don’t actually possess any knowledge in this particular example because it is a COINCIDENCE that my belief is true, or to be more specific, it’s “true-ness” is not related to my justification (i.e. that I see her on the couch). It’s “true-ness” is based on the coincidence of her actual presence behind the couch and thus meeting the requirements of “being in the living room” (albeit without me knowing it).
How do we get around this dilemma, and can we do so such that we at least partially (if not completely) preserve Plato’s requirements? One way to solve this dilemma would be to remove the element of COINCIDENCE entirely. If we change our definition to:
Knowledge = Justified Non-Coincidentally True Belief
then we have solved the problem presented by Gettier. Removing this element of coincidence in order to properly define knowledge (or at least in order to define it more appropriately) has been accomplished in one way or another by various responses to the “Gettier Problem” including that of Goldman’s “causal theory”, Lehrer-Paxson’s “defeasibility condition“, and many others which I won’t discuss in detail at this time, but the links provided will show a basic summary of some of those responses. Overall, I think it’s an important distinction and a necessary one to make (i.e. non-coincidence) in order to properly define knowledge.
Alternatively, another response to the “Gettier Problem” is to say that while it appears that I didn’t actually possess true knowledge in the example illustrated above (because of the coincidence), it could be due to the fact that I never actually had sufficient justification to begin with as I failed to consider the possibility that I was being deceived on a number of possible levels with varying significance (e.g. hologram, brain in a vat, etc.). This lack of justification would mean that I’ve failed to satisfy the Platonic definition of knowledge and thus no “Gettier Problem” would exist in that case. However, I think it’s more pragmatic to trust our senses on a basic level and say that “seeing my wife in the living room” (regardless of whether or not it may have actually been a hologram I was seeing and thus wasn’t “real”) warrants a high level of justification for all practical purposes. After all, Plato never seemed to have specified how “justified” something needs to be before it meets the requirement necessary to call it knowledge. Had Plato considered the possibility of the “Brain in a vat” scenario (or a related scenario more befitting to the century he lived in), would he have said that no justification exists and thus no knowledge exists? I highly doubt it, since it seems like an exercise of futility to define “knowledge” at all if it was assumed to be something non-existent or unattainable. Lastly, if one’s belief justification is in fact based on hierarchical patterns observed in nature, then the more detailed or thoroughly defined those patterns are (i.e. the more scientific/empirical data, models, and explanatory power that are available from them), the more justified my beliefs will be and the more useful that knowledge will be.
In summary, I think that knowledge is nothing more than an attribute that we possess solely based on an understanding of and adherence to definitions and thus is entirely compatible with the BIV scenario. Definitions do not require a metaphysical explanation nor do they need to apply in any way to anything outside of our reality (i.e. metaphysically). They are nothing more than human constructs. As long as we are careful in how we define terms, adhere to those definitions without contradiction and understand those definitions that we create (through the use of hierarchical patterns observed in nature), then we have the ability to possess knowledge — regardless of if we are nothing more than BIVs. The only knowledge that we can claim to know must be “justified non-coincidentally true belief” (to use my modified version of the Platonic definition, thus accounting for the “Gettier Problem”). The most significant realization here is that the only knowledge that we can PROVE to know is based on and limited by the definitions we create and our use of them within our claims of knowledge. This proven knowledge would include but NOT be limited to Descartes’ famous claim “cogito ergo sum” (i.e. “I think therefore I am”), which in much of Western Philosophy is believed to be the foundation for all knowledge. I disagree that this is the foundation for all knowledge, as I believe the foundation to be based on the understanding and proper use of definitions and nothing more. For Descartes, he had to define (or use a commonly accepted definition of) every word in that claim including what “I” is, what “think” is, what “am” is, etc. Did his definition of “I” include the unconscious portion of his self or only the conscious portion? I’m aware that Freudian theory didn’t exist at the time but my question still stands. Personally I would define “I” as the culmination of both our conscious and unconscious minds. Yet, the only “thinking” that we’re aware of is conscious (which I would call the “me”), so based on my definitions of “I” and “think”, I would disagree with his claim, because my unconscious mind doesn’t “think” at all (and even if it did, the conscious mind doesn’t perceive it so I wouldn’t say “I think” when referring to my unconscious mind). When I am thinking, I am conscious of those thoughts. So there are potential problems that exist with Descartes’ assertion depending on definitions he chooses for the words used in that assertion. This further illustrates the importance of definitions on a foundational level, and the hierarchical patterns observed in nature that led to those definitions.
My views of knowledge may seem to be unacceptable or incomplete, but these are the conclusions I’ve arrived at in order to avoid the “Gettier Problem” as well as to remain compatible with the BIV scenario or any other metaphysical possibilities that we have no way of falsifying. While I’m aware that anything discussed about the metaphysical is seen by some philosophers to be completely and utterly pointless, my goal in making the definition of knowledge compatible with the BIV scenario is merely to illustrate that if knowledge exists in both “worlds” (and our world is nothing but a simulation), then the only knowledge we can prove has to be based on definitions — which is a human construct based on hierarchical patterns observed in our reality.
The Open Mind 
Congratulations on your shiny new blog.
I’m inclined to say that’s the wrong approach. I go by Wittgenstein’s “meaning is use.”
When I was young, “gay” meant something like “cheerful”. Now it has a very different meaning. So how did the meaning change? This came from an extensive grass roots campaign by the gay community, that resulting in a widespread change of usage. I don’t think you can alter meanings in any other way.
Within science, we often do precisely define our terminology. But even that is consistent with Wittgenstein, in the sense that using precise definitions is part of normal scientific usage, well supported by the traditions of science. By contrast, the traditions of philosophy support rather sloppy usage and vague indeterminate meanings.
There are some statements that I tentatively hold to be true, yet that I question from time to time. There are other statements that I believe quite strongly, and you would not easily get me to question them.
My point: There are degrees of belief, but this is omitted in the JTB characterization of knowledge.
As for “true” – there are serious problems with a correspondence theory of truth. It is unworkable. And a coherence theory of truth is probably no better. Since your topic is BiV, what could correspondence even mean to a BiV?
I am not actually trying to start a debate here. I’m just offering some things to ponder, while giving the comments section of your blog a jump start.
Thanks for the congrats. I figured if I enjoy writing as much as I do, why limit that writing to comments when I can create some posts of my own as well. I haven’t learned all the functions/options yet, in terms of block quotes, etc., so bear with me.
I disagree with Wittgenstein in that regard. His assertion that even without a definition, we still successfully use words misses an important point in my opinion. Mainly, any word that is 100% successfully used has to have a definition to describe it’s meaning. This does not mean that I have to read a definition in order to understand a word in some way. It just implies that our inference of the meaning of any particular word which we did not create ourselves can only be verified to be correct with definitions. In the case of knowledge, we need this verification if we are to remain on solid ground.
I don’t know how you came to the conclusion that this is the only way to alter meanings of words. We can alter them anytime we want to by defining the word differently. In the case of “gay”, it now has two different meanings which are synonymous with the words “happy” and “homosexual”.
The more sloppy that usage is, the less likely we can verify if an idea is communicated successfully, let alone if it is logically sound at all. If one is merely a philosopher that only contemplates inside his head then this may be alright as that philosopher knows what he or she means by the words used. However, when communicating an idea to anyone else, that sloppiness has to be minimized if not eliminated in order to communicate effectively.
I think that the only thing that varies is your conscious view of the justification for believing it. Either you believe a proposition is true or you believe it is not true (or you have no belief on the proposition). I doubt you believe some of your beliefs are more or less true than others. They are either true or false. The varying “degrees of belief” that you speak of are (in my opinion) really just a variation in the self-evident justification for the belief, and justification is a main component of the JTB characterization of knowledge.
I agree that there is a problem with a correspondence theory of truth, but again my entire contention is that the only truth (more specifically knowledge) one can attain is based on definitions. This circumvents those serious problems in my opinion.
I’m not sure what you’re asking here exactly. “Correspondence” would mean exactly what we currently define it to mean. To the mad scientist (or agent responsible for the simulation), we could say that if they referred to the “reality” that the BIV is experiencing internally, they may say that correspondence would be defined the same way within that virtual “reality” as the scientist would define it in their reality. If that didn’t answer your question adequately, you may have to be more specific in what you’re asking.
The problem with that view of language, is that we can never know what the definition is. That’s roughly Wittgenstein’s argument on the impossibility of following a rule.
But maybe meanings are not the kind of things that could ever be inferred.
If I define a word differently, then that word becomes useless for communication, unless I can persuade lots of other people to also define it differently.
My question was not about what correspondence means to the scientist. Rather, it was about what correspondence means to the BiV itself (to whatever entity is experiencing the simulation).
Wittgenstein doesn’t appear to offer any alternatives to understanding meanings of words. If we don’t use definitions, then how can we verify that the word’s meaning has been inferred correctly? For that matter, how can we decide what a word means if there are no known meanings of words? Do we just assume so? Why not just abandon all definitions and dictionaries then?
In that case, I think that words would be meaningless if their meaning could never be inferred. There has to be some idea of what a word is supposed to mean and the only way that happens (that I can see) is through the agent’s inference of that meaning. The only reliable way of verifying (at all) if that inference is correct is through the use of definitions.
Well they don’t have to define it differently, rather they only have to see what your definition is to know what you mean by using the word. It doesn’t become useless if you share the definition of that word with those you communicate with (we certainly don’t need to communicate novel words/meanings with everyone in the world – only when it is necessary which is not very often). Once they are aware of it, then communication is successful. Words come and go all the time and there has to be some way for those words to change. It starts with one person that creates a new word (or creates an alternate meaning for an existing word), then eventually if others agree to the definition and/or it is propagated well, it will become a mainstream word. There are plenty of words that are in the dictionary that most people don’t use, let alone nomenclature in certain esoteric groups in science, etc. It’s only important that the finite number of people in the discussion are aware of the definitions of words used.
We would be the BIVs in this scenario so it would mean whatever we currently think it means.
We obviously have very different ideas about language.
What is your answer to my question? If we don’t use definitions, then how can we verify that the word’s meaning has been inferred correctly?
How do we know (or how can we know best) if someone understands what a word is if they can’t explain their understanding through words?
Don’t get me wrong. I do think that definitions (and thus words as well) do have limits and it can be a difficult task assigning certain words a specific all-encompassing definition. I also am aware that in order to understand a definition, one has to break apart each word within the definition and know it’s constituent definitions (and their respective meanings). However, words that have a complex meaning can be broken down into a set of words where each constituent word has a more simple meaning and thus definitions come into play for not only better understanding complex words by breaking them down into less complex words, but also by having some external means to at least help verify inferred meanings. If you’d like to offer an alternative (other than definitions) to verify (or at least most effectively verify) the inferred understanding or meaning of a word, I’d like to hear it.
In the field of developmental psychology or even evolutionary psychology for that matter, one focus is on how we learn our first words (as infants ontogenetically and as early human ancestors phylogenetically) since we obviously can’t use words to define and understand those new word’s meanings. It’s clear that as pre-language beings we’d have to observe non-verbal cues like pointing to an object while uttering a sound and then associating the two together (the sound being a symbol for representing what was pointed at). We have to possess several mental concepts including the idea of objects (and their permanence in most cases), the ability to build associations, and most importantly we need to possess the understanding of symbolic representation. Once the association is in place between a word and an object, then the vocabulary can grow with new associations in combination with the existing vocabulary to help give it context. Once we’ve done this to a point where a plethora of simple words have been learned, then the addition of new complex words/concepts which may be difficult if not impossible to learn through that elementary and fundamental association process, can only be learned (with the meaning best verified) through definitions. That is my main point here. I’m aware that the building blocks of language had to involve non-linguistic components, but once the building blocks are in place then linguistic components need to be employed for learning (and verifying) complex concepts and words.
I am disagreeing with the question. Language doesn’t work that way.
Meaning is subjective. It comes from within, not by inference from without.
A language speaker can tell whether he/she is using words appropriately, by observing the reactions of others involved in the discussion. We learn appropriate usage. And that is “appropriate” usage, not “correct” usage.
I agree with you 100% that meaning is subjective and comes from within. I don’t argue that. However, my question was addressing the verification component of inference. How can we have any idea what subjective meaning another person assigns to a word? You’ve said that we can observe the reactions of others based on the speaker’s usage, but that is a fairly crude method which is less repeatable, not quantifiable nor qualitative in any controlled way. Different groups or members of that audience could be focusing on different parts of my speech and not notice subtleties that can dramatically change the meaning of what the speaker is saying. I could have small changes in my internal subjective meaning I assign to a word or concept, and people may miss it in speech. If however, I explained what I think something means via a definition (along with monitoring their reaction to my speech), there is at least a common linguistic platform with which to describe that meaning. In other words, it’s as close as we can get to a quantifiable/qualitative/repeatable metric for verifying inference of meaning. Meaning can still be subjective and internal, but that doesn’t mean that there aren’t some tools we can use (e.g. definitions) to help establish more objective verification of that meaning. That is — some linguistic way of at least partially explaining what’s going on inside the head of the speaker.
As best I can tell, the terms “inference” and “verification” apply only to the objective. They are not appropriate terms for subjective questions.
You are attempting to objectify the subjective. I don’t know of any reason to suppose that should be possible.
Yet it seems to be adequate for being able to use language in a way that we can effectively communicate.
This is true. We deal with that by engaging in dialog. And our disagreements in that dialog help bring our attention to what was not adequately communicated. Then, as part of the dialog, we attempt to deal with that, though not always successfully.
When I say “verify” I’m not implying 100% verification (which you are assuming by associating it exclusively with the objective). I am only implying that definitions will help us achieve more verification (evidence in favor of). We can still use an objective lens to look at things that are subjective to provide some kind of weight to them. For example I could ask people to tell me what they usually mean by the phrase “I’m feeling ok” using a scale of 1 to 10, where ‘1’ is neutral/normal and ’10’ is the best they’ve ever felt. We will still never know objectively what they mean, but we will have a better sense of trying to quantify it by comparing it numerically with other subjective feelings. Does this make sense? The answer that the person gives me, whether it’s a 1 or a 10 is not meaningless information. It would be a way to quantify their inferred meaning of the word “ok” relative to other emotions they’ve experienced. Also, the term “inference” doesn’t apply only to the objective. I’m not sure where you gathered that idea. Their are implications of words used (by the sender), and their are inferences of words used (by the receiver). They are complementary to one another. The concepts or ideas that a word “suggests” to a person is that person’s inference of the word. It is subjective.
I don’t think it’s adequate at all (not for complex language anyways). Watching the reactions of others is not enough to demonstrate that ideas have been communicated effectively. Have you ever had someone smile and nod while you were talking to them, only to find out later that they had no idea what you were talking about? It’s happened to me. If this has never happened to you, then consider yourself lucky (it would explain why you have the opinion that you have on this issue). If definitions of words are spelled out prior to the speech, THEN the reactions of others will have more weight because they have been given some common platform to work with. Have you ever noticed that some books or articles with complex nomenclature have a glossary? Why is that do you suppose? In the case of reading something, there is nobody to see your reaction to the words read. Definitions provide a way to HELP verify (though not 100%) effective communication in any medium (written or spoken). Any ambiguities (many of them anyways) as well as words that have complex meanings can be explained in some way prior to the speech.
Disagreements in that dialog may not precipitate if certain elements in the speech aren’t questioned. Beyond that, definitions help to further deal with that issue (in a more objective way), especially when engaging in dialog isn’t an option in certain settings.
Okay, that’s getting much closer to a realistic account of language.