• TheMadFool
    8.6k
    According to the Internet Encyclopedia Of Philosophy, the most general form of The Problem Of The Criterion is as follows:

    1. Which propositions are true/knowledge? [Instances of truth/knowldege]

    2. How can we tell which propositions are true/knowledge? [Definition of truth/knowledge]

    Other ideas relevant to the discussion can be found on the same page.

    The Problem Of The Criterion can be formulated with two propositions given below:

    A. We can't know instances of truth without having a definition of truth/knowledge [methodism]

    B. We can't have a definition of truth without knowing instances of truth/knowledge. [particularism]

    The argument is that A and B form a vicious circle: To identify instances of truth/knowledge, we need to first formulate a definition of truth/knowledge but to formulate a definition of truth/knowledge we must first possess instances of truth/knowledge and the vicious circle is complete and we're helplessly trapped inside it...or so it seems.

    The statement A makes complete sense for how can we find a thing without knowing what that thing is? For instance, I don't know what "parsing" means and I can tell you with full confidence that I wouldn't be able to identify instances of "parsing" (By the way I would be grateful if someone could edify me about what "parsing" means. Thanks in advance.)

    What about statement B? Does it make sense?

    Is it possible to recognize dogs/truth/knowledge without knowing what dogs/truth/knowledge are/is? This smacks of a contradiction: I don't know what dogs/truth/knowledge are/is (I don't have a definition of dogs/truth/knowledge) & I know what dogs/truth/knowledge are/is (I can identify dogs/truth/knowledge). Since statement B is/entails a contradiction, it must be false.

    Ergo, particularism is untenable and the way to go is methodism.
  • Possibility
    1.9k
    Without reading the IEP entry yet, just going on your description here...

    I have a memory of my daughter learning to speak that comes to mind here. Aged somewhere around 18 months old, she would point out the window of the car as we drove along and say ‘bah’. It took us a number of instances of this to recognise that she was pointing to puddles and making a connection between them and her nightly bath. She was using a language concept ‘bah’ to describe instances of water. So, did a definition develop later for her, or is definition simply a summary of instances?

    I think the circular nature of this process is important, but I also think the language concept is a vital piece that seems to be overlooked. It is possible for a child to correctly label an instance of ‘dog’ without a definition of what dogs are - with nothing more than a fuzzy language concept formed from a summary of previous instances. Does this mean they know instances of ‘dog’? The main difference between a child not knowing what ‘dog’ means and you not knowing what ‘parsing’ means may simply be the courage to risk being wrong in identifying an instance.

    I think there’s also a distinction to be made here between the possibility of truth and the potentiality of knowledge.
  • TheMadFool
    8.6k
    I think your take on the matter has some merit but it breaks down on closer scrutiny. Your daughter being able to identify dogs (knowing dogs) without a definition of dogs (not knowing dogs) is an through and through contradiction as I've described in the OP - it's an impossibility! How can I find something when I don't know what that something is? Don't take my word for it...try it out for yourself if you don't believe me.
  • baker
    568
    How can I find something when I don't know what that something is?TheMadFool
    The process of "discovering" truth is simultaneously deductive and inductive.
    Assuming something to be true makes us able to see the supporting evidence.

    Truthing is circular like that.
  • baker
    568
    Truthing is circular like that.baker
    Unlike parsing, heh.

    images?q=tbn:ANd9GcRnttiLP_UYWOiIwdCrrp68ErXbEse2cr0Www&usqp=CAU
  • simeonz
    208
    The definition of truth is complex and it will be useful to tackle it separately. It raises questions beyond the above dichotomy, such as to what does it apply, whether it is a matter of accuracy of our representations or efficacy of our planned courses of action. Rather confounding issues.

    So, I will rephrase the statements as they apply to classification:

    A. We can't discern members of a class without having criterion that identifies their features. [methodism]

    B. We can't specify criterion for the members of concrete class without being presented with examples that we can extrapolate to a definition. [particularism]

    In computer science, this is the clustering problem. Given unclassified samples, we are tasked to break them into spatially contiguous groups, while simultaneously identifying which grouping produces the least internal sparsity. The solution is computationally laborious, intractable when perfectly optimal, but its approximations are reasonably efficient. The applications sometimes require human discretion (supervision as it is called) for domain-specific considerations, but automatic execution is not impossible. Solving such a problem presupposes an already established feature space, so the techniques may have to be combined with feature extraction / dimensionality reduction step. This is the step in which you conclude which qualities are most distinguishing for your samples. For the human cognition, the task may require more sophisticated mechanism, but the problem should be otherwise similar. We discover objects and situations which have closeness in some regard and define the categories that naturally emerge as the most effective classifiers.

    Having derived classes for objects or situations, and having the faculty to introduce conjunctions (thus universals), disjunctions (thus existentials), complements, modalities, etc, a human being may produce a reasonable amount of propositions. I am not sure whether thinking operates with such formality and I am not sure whether we introduce logic into the picture in precisely this way, but it offers one path to the ability of autonomous mental synthesis of statements.

    Feature extraction
    Dimensionality reduction
    Cluster analysis

    I am not sure about which kind of parsing you are referring to, but in computer science, and I believe, also in linguistics, it is the process of deciding how to split a text into its syntactical constituents and label them with their syntactical categories (ascribe meta-information to each region of letters and symbols, metadata as it is called in computing).
  • TheMadFool
    8.6k
    To all

    That there are different definitions of truth (correspondence, paragmatic, coherent, etc.) is suggestive...hints at some degree of arbitrariness...something I referred to in the OP.

    If truth were abstracted from instances of truth this wouldn't be the case for then that which can be described as the form (Plato?) of truth would be constant, precluding, in my humble opinion, variety in the definition of truth.
  • Possibility
    1.9k
    I get what you’re trying to say, but I think you’re oversimplifying.

    Being able to identify an instance of dog is not the same as knowing dogs. By the same token, any definition of ‘dog’ is not necessarily as exact as it claims to be. So, while knowing dogs without knowing dogs IS an obvious contradiction, I maintain that it IS possible to find or identify an instance of something without knowing definitively prior what that something is - until you find it, that is. In fact, it appears to me that knowing an instance and knowing a definition may indeed be one and the same process.

    That there are different definitions of truth (correspondence, paragmatic, coherent, etc.) is suggestive...hints at some degree of arbitrariness...something I referred to in the OP.

    If truth were abstracted from instances of truth this wouldn't be the case for then that which can be described as the form (Plato?) of truth would be constant, precluding, in my humble opinion, variety in the definition of truth.
    TheMadFool

    The more instances (limited definitions) of truth we incorporate, the closer we may get to a broad ‘definition’ of truth. But in all honesty, I think truth would be formless as such - any possible ‘definition’ of truth is necessarily inclusive of its relation to what is not true.

    We keep returning to a contradiction, and we keep rejecting it, convinced there must be some other, more logical structure underlying it all...
  • simeonz
    208
    That there are different definitions of truth (correspondence, paragmatic, coherent, etc.) is suggestive...hints at some degree of arbitrariness...something I referred to in the OP.TheMadFool
    I was concerned with the possibility of solving the issue as specified and wanted to propose that we break it down into aspects of smaller complexity. I may have come off as assuming some lack of familiarity with the intricacies that are involved.

    If truth were abstracted from instances of truth this wouldn't be the case for then that which can be described as the form (Plato?) of truth would be constant, precluding, in my humble opinion, variety in the definition of truth.TheMadFool
    It still seems to me that the criterion is not about establishing unambiguously the truth of propositions, but of designing descriptions for a proposition that matches particular experience. The problem does elicit however many considerations involved when matching the descriptions between individuals through exemplification.

    Lets consider correspondence truth first. The assertion that an object is an apple is just a statement for association with some class of objects, and different people may have different notions in this regard. A child thinks of an "apple" and so does an agricultural botanist. First, the child may have had limited experience with apples in relation to its interactions and needs. The object have been measured as fitting in the palm of your hand, the color reddish with sparse nuances of green, the taste, mostly acidic sugary or just sugary. A botanist, in contrast, knows apples of all varieties, ripe and unripe, green and red, bigger and smaller fruits, etc. It will measure the sizes in widely recognized units, not in proportion to body parts. Nonetheless, both will associate their descriptions with apple. The ambiguity stems from matching the statement and not the interpretation. (Edit: What I mean is, the statement is considered the same, even though, both parties would agree empirically that their interpretations do not match.) The botanist is at least uncontested authority in this case, which is not always a clearly defined role, and hence, the child can be informed of the class's variety through literature, expressed in references to other words, or through other examples. But people will ultimately never base their vocabulary in the same set of direct experiences. (*)

    I will use the mechanism of clustering to explicate some of the issues involved. It is not the value of the proposition which is ambiguous. The indefiniteness of the statement is the problem, which is inherent in the way in which the description is derived. The clustering algorithms will produce a solution based on the quality and variety of the sample (the objects) and the desired crudeness. Those factors will vary for different people. The feature set used as relevant discriminator will also differ. And the truth will not be binary, but will have a degree of certainty, with clear matches and some borderline cases. The confidence when dealing with uncertainty will differ, based on the needs and attitudes of the individuals.

    The same logic applies to pragmatic truth, but instead of representations, it is dealing with intents and applications. For example, what is right for the poor is not right for the rich, what is right for the elderly is not right for the young, etc. Different people will associate the truth of certain statements with different social clusters or conditions, and thus will interpret them into different propositions.

    *: This example is easily rectifiable error of conceptualization and communication due to limited observations of one party, which also shows that representational truth is inherently easier to verify. This is why it is so central to science. As you suggested, one can talk about some sense of objectivity there, although the question of bringing all experiences in alignment remains problematic. But pragmatic truth, a first-order phenomenon of cognition in my opinion, has much less sense of objectivity. It also affects debates of proposed representations, because the very criteria for making a proposition canonical can be affected. For example, some physicists would argue that a theory of should be mathematically minimal, while others will argue that it has to confer at least some explanative value.

    P.S. In summary, I think the circularity is not about ultimate truth, it is about finding effective taxonomy of the available knowledge. Even so, the process is one of continuous refinement (for example, exploration of the facts by science and redefining the concepts after some cluster of information is supplemented with new data), distributing the relevant information to all authorities on the subject, and arbitrating between attitudes and values that affect the perception of the explanative scope of the evidence, its significance relative to the individual cognitive bias, culture, social strata.
  • creativesoul
    9.7k
    A. We can't know instances of truth without having a definition of truth/knowledge [methodism]

    B. We can't have a definition of truth without knowing instances of truth/knowledge. [particularism]
    TheMadFool

    I disagree.

    Language users can know when some statements are true or not long before they have a definition of truth or knowledge. That much is easily proven.
  • frank
    6.4k
    The statement A makes complete sense for how can we find a thing without knowing what that thing is?TheMadFool

    This is Meno's paradox. Plato uses it to show that some knowledge is innate (that's one way to put it).
  • TheMadFool
    8.6k
    Meno's paradox.frank

    :up: :ok:

    Initially I was wondering whether Possibility's daughter's ability to identify dogs had something to do with innate knowledge but the matter is much simpler than that. Pointing to dogs and uttering the word "dog" is an act of providing instances to the audience (here Possibility's daughter) and if that's all that's being done, leaving the audience to figure out what the word "dog" means i.e. it's the audience's job to abstract the essence of a dog from the instances provided. It appears this is a valid method of defining words. That's that.

    Defining truth may be similar too. We do a systematic survey of propositions and sort them based on different attributes and decide that propositions with such and such attributes (whatever they maybe) should be called true propositions and absent these attributes are not true.

    However, there's an, for lack of a better word, intuition albeit vague as far as I can tell that truth has to be something specific i.e. there are constraints on what truth can be. The thorough study of the atrributes of porpositions don't result in truth being defined based on just any constellation of attributes. To the contrary, we're drawn to certain groups of attributes (correspondence, coherence, pragmatic, etc) - it feels natural to define truth in these terms - and this I consider as an indication of a preconceived, how shall I put it, idea of what truth should be.

    In other words, it may look like we're trying to abstract a definition of truth from instances of truth, from a careful analysis of propositions but in fact we already possess a definition of truth and are simply looking for propositions that match that definition. That is to say that, at least on the matter of the definition of truth if not dogs the impression that we get of examining propositions so that we may extract the essence/form of truth is an illusion.

    What say you?
  • Possibility
    1.9k
    Initially I was wondering whether Possibility's daughter's ability to identify dogs had something to do with innate knowledge but the matter is much simpler than that. Pointing to dogs and uttering the word "dog" is an act of providing instances to the audience (here Possibility's daughter) and if that's all that's being done, leaving the audience to figure out what the word "dog" means i.e. it's the audience's job to abstract the essence of a dog from the instances provided. It appears this is a valid method of defining words. That's that.TheMadFool

    Well, my take on this is not essentialist, so I don’t see it as ‘abstracting the essence of dog from the instances’, but as abstracting recurring patterns in qualitative relations. This will always be ‘fuzzy’ to a certain extent - a definition seems to be just a linguistically-structured summary or reduction of these patterns.

    Defining truth may be similar too. We do a systematic survey of propositions and sort them based on different attributes and decide that propositions with such and such attributes (whatever they maybe) should be called true propositions and absent these attributes are not true.

    However, there's an, for lack of a better word, intuition albeit vague as far as I can tell that truth has to be something specific i.e. there are constraints on what truth can be. The thorough study of the atrributes of porpositions don't result in truth being defined based on just any constellation of attributes. To the contrary, we're drawn to certain groups of attributes (correspondence, coherence, pragmatic, etc) - it feels natural to define truth in these terms - and this I consider as an indication of a preconceived, how shall I put it, idea of what truth should be.

    In other words, it may look like we're trying to abstract a definition of truth from instances of truth, from a careful analysis of propositions but in fact we already possess a definition of truth and are simply looking for propositions that match that definition. That is to say that, at least on the matter of the definition of truth if not dogs the impression that we get of examining propositions so that we may extract the essence/form of truth is an illusion.
    TheMadFool

    Here’s a better word than intuition: assumption.

    To define something - to state or describe exactly its nature, scope or meaning; to mark out its boundary or limits - is a reductionist methodology that discards qualitative variability or ‘fuzziness’ in the information we have about that something.

    It feels more comfortable to ignore, isolate or exclude a relation to truth that lacks sufficient attributes to be positioned with certainty on this side of an arbitrarily-drawn true/false dichotomy. The idea of truth includes an understanding why we feel so uncomfortable with this uncertainty. The idea of what truth should be, however, excludes this relation to what is possibly but not certainly true.

    Are you suggesting here that a definition of truth is a priori analytical knowledge?
  • TheMadFool
    8.6k
    not essentialistPossibility

    What's the alternative? Anything goes? So, for instance, a dog could be defined in terms of non-essential features like fur, claws, ears, eyes, tail, fangs but then...event cats, bears, tigers have these and then every one of these essentialism-based categories would be dogs. Do you want to go down that road? I could be mistaken of course and that's where you come in I guess.

    fuzzinessPossibility

    Nec caput nec pedes. Can you clear the matter up for me? I don't see the relevance of fuzziness to The Problem Of The Criterion. For my money, the issue of vagueness comes much much later - after we've settled the matter of what truth means and which statements are true. Even if truth is a fuzzy concept there have to be propositions that are clear-cut truths.
  • simeonz
    208
    Pointing to dogs and uttering the word "dog" is an act of providing instances to the audience (here Possibility's daughter) and if that's all that's being done, leaving the audience to figure out what the word "dog" means i.e. it's the audience's job to abstract the essence of a dog from the instances provided.TheMadFool
    This will always be ‘fuzzy’ to a certain extent - a definition seems to be just a linguistically-structured summary or reduction of these patterns.Possibility
    To define something - to state or describe exactly its nature, scope or meaning; to mark out its boundary or limits - is a reductionist methodology that discards qualitative variability or ‘fuzziness’ in the information we have about that something.Possibility

    I am not saying that propositions are inferred from example instances for each specific proposition, but by dividing the instances of experience into suitable classes. Only when it comes to synchronizing the vocabulary and learning through normative (compulsive) and authoritative (convincing) sources, the classes are acquired second-hand or through especially representative examples.

    I am not saying that someone will point dogs to a child until it is coerced to learn the concept dogmatically. I am saying that a child will see several dogs and cats, and will be able to use its own discretionary capacity to derive that "this is one thing, and that is another thing". Its mother might convey to her/him, that we call the first a dog and the second a cat, but the conceptualization of the species is already present. And I pointed out that we are not merely speculating about the possibility of such classification, but that there are ways to specify the problem precisely and to solve it with available techniques.

    Obviously, there is fuzziness, because the subject has influence on the parameters of the task. A child might decide that koalas are a kind of bear. Which they are not. But for a child's purposes the distinction is inconsequential. That is why the intentions and needs will factor themselves in. The child will only rectify its concept under coercion (they might be simultaneously provided convincing justification), because it has no need to do so on its own. Many parameters will depend on the subject - the completeness of the description, i.e. known features, of the instances, the crudeness of the classes, or clusters, the total sample collection (have they seen enough canines, felines, bears, marsupials, aliens, sledgehammers, quantities of things, lengths of things), and how they metricize the feature space (i.e. how they define the distance between objects/situations/qualifiers in consideration of their biases, attachments, goals, etc)
  • SophistiCat
    1.6k
    I am not saying that someone will point dogs to a child until it is coerced to learn the concept dogmatically.simeonz

    I see what you did here
  • simeonz
    208
    I see what you did hereSophistiCat
    I wouldn't have noticed. It happened accidentally.
  • frank
    6.4k
    Initially I was wondering whether Possibility's daughter's ability to identify dogs had something to do with innate knowledgeTheMadFool

    Face recognition is innate. Dogs have forward facing eyes like a human, so dogs should be stand out to babies.
  • Possibility
    1.9k
    not essentialist
    — Possibility

    What's the alternative? Anything goes? So, for instance, a dog could be defined in terms of non-essential features like fur, claws, ears, eyes, tail, fangs but then...event cats, bears, tigers have these and then every one of these essentialism-based categories would be dogs. Do you want to go down that road? I could be mistaken of course and that's where you come in I guess.
    TheMadFool

    The concept ‘dog’ is constructed in our minds with the help of language in relation to instances. So, a ‘dog’ may be initially understood in terms of a relational structure of shapes, size, sound, texture, etc. - depending on whether those early instances are a family pet, pictures in a book, or sounds from next door. This is how my daughter initially understood ‘bath’ to describe bodies of water. From there, she soon realised that ‘bath’ referred to more specific instances of ‘water’. She may also see another furry creature with pointed ears, four legs and a tail and say ‘dog’ - only to be gently corrected with ‘cat’. Remember ‘Monsters Inc’, when the little girl calls the big furry monster ‘Kitty’? It’s not about essential features, but about recognising patterns in qualitative relational structures.

    Nec caput nec pedes. Can you clear the matter up for me? I don't see the relevance of fuzziness to The Problem Of The Criterion. For my money, the issue of vagueness comes much much later - after we've settled the matter of what truth means and which statements are true. Even if truth is a fuzzy concept there have to be propositions that are clear-cut truths.TheMadFool

    Why do there have to be propositions that are clear-cut truths? In order to think, speak or act with a degree of certainty or confidence in what is, for all intents and purposes, a prediction. We haven’t settled the matter - we’ve constructed a prediction, which we’ve then defined in a summary of past instances. The accuracy of this definition is temporary: fragile and fleeting from the moment it’s proposed. Hence the fuzziness of the concept.
  • TheMadFool
    8.6k
    It’s not about essential features, but about recognising patterns in qualitative relational structures.Possibility

    I beg to differ. In the absence of essences to dogs or whatever else is the topic, there can be no further discussion. Can you tell me what "dog" means? I'm supremely confident, as out of character as that is, that you'll be listing a set of essential features.

    fuzzinessPossibility

    Why are we discussing predictions?
bold
italic
underline
strike
code
quote
ulist
image
url
mention
reveal
youtube
tweet
Add a Comment