Friday, December 19, 2014

We Don't Evade the IRS: A Nietzschian Criticism of Scientific Reliabilism

       The epistemic foundations of science rest, in part, on presumptions concerning language. One such assumption is that language is literal, that we describe things or that statements and utterances correspond to physical objects, or events. There is obviously a practical motivation guiding this as the world easily becomes more intellectually digestible: those things in the “sky” are “clouds”; there is “concrete” underneath my “feet”. Having a fixed system of terms towards particular things, and their properties, allows us to feel familiar with our environment by organizing things in a manner that allows for intellectual development. The prose of Dr. Seuss and Vonnegut are both thought to deal in fictions, while science is about reality. This dichotomy keeps scientific law intact in that there are statements that just don't adhere to reality. The child who says “the sky is purple,” is thought to be imaginative mainly for the manner in which she deviates from our fixed truths: on a clear day “the sky is blue.” Nonetheless, in time, the child will be made to make the correct utterances, and for the person who remains imaginative there will be certain punishments. We will think of her as immature, stupid, or having some sort of disability. Having the proper sight requires taking words seriously, as though they are glued to objects. The more we take words literally, the more we suffocate our creative instincts.
  One area of thought, Logical Positivism, views statements that don't pertain to physically verifiable conditions as meaningless. Apart from this main tenet, there is no fixed ideology for logical positivism, even for those involved with the movement, but what seems common is a concern for scientific methodology and its impact on society. Ideas, such as phenomenal information, positivists think are typical of metaphysics in that claims within the tradition offer no means of physical measurement or testing. In order to combat these claims, logical positivists developed a linguistic approach to science aimed at establishing a criteria for meaningful claims. The meaningfulness of a claim depends on whether the claim in question has the ability to be tested. Positivists view science as an empirical practice governed by formal sentential relations, which are predicated on first order logic.
  The positivist view was unaccepted not only by metaphysicians, but also by those within the scientific community. In particular, Bogen and Woodward offer the idea of scientific reliability to challenge the notion that evidence supports a hypothesis due to a formal relation between sentences. The term 'reliability' pertains to the quality of evidence involved in scientific experiments, which is determined by the dependability of the processes that produced them. They propose that [it is better to focus directly on the data and the processes that generate them],1 in order to determine whether something is epistemically relevant. This contrasts greatly with positivists who generally think that what we test are statements. Reliabilism maintains that this cannot be the case seeing as some pieces of evidence are extra-linguistic, that they are not terms or statements, rather, they are actual physical aspects of the world. Bogen and Woodward argue that the relationship between evidence, which they refer to as data, and phenomena is not sentential, rather, it is causal.
Each of these views on knowledge and language inspired me to reflect on Nietzsche, seeing as I think he would have very specific criticisms toward both. Though they are not the only two theories that comment on what we need, and don't need, language for, I'm intrigued by this notion of extra-linguistic objects, and how we are supposed to epistemically interact with them. If these objects are not linguistic, then how am I to intellectually absorb them? As for positivism, I'm astonished at the overabundance of weight that is placed on the role of language in epistemic affairs, and how they feel able to belittle commentary outside of the scientific realm. I consider how Nietzsche believes all language to be metaphoric, and that the statements we construct concerning the reality of our affairs is a creative gesture, not a logical one. Though we overlook the metaphoric nature of language, we cannot avoid using it if we want to be educated on the world.
Beginning with logical positivism, and Carnap's rendition of it into hypothetico-deductivism, I will discuss the principle of verification, which distinguishes meaningful statements, those statements that say something about the world, from meaningless ones. From here we will see how positivists organized terms into two different groups: observational and theoretical terms. This allows for restrictions concerning what types of claims can be said about the world. The section following this details Bogen and Woodward's account of reliabilism that is pitted against positivism. Here I examine the difference between data and observational statements, and which should count as evidence. Bogen and Woodward claim that it is not data that theories explain, but it is phenomena. They emphasize that science advances through a rigorous testing of data-generating procedures, and not through any sort of relation between sentences.
From my interpretation of Nietzsche, I will attempt to show that reliabilism is not much different than any doctrine of formal sentential relations. The scientific procedures Bogen and Woodward argue for amount to nothing more than additional sentences about data-generating procedures. Despite language being metaphoric in nature, Nietzsche feels that when it comes to matters of intelligibility, language, and having a relation between sentences, is indispensable. As such, there are no extra-linguistic items. This is not to defend positivism, as Nietzsche does not think language reports on the world. When it comes to using language to describe, or denote the things we observe, words and statements do not correspond to things in the world. Statements creatively relate only to other statements, and this closed system, Nietzsche believes, is the foundation to knowledge. By undermining these scientific notions, I hope to reveal no one statement, or series of which, should be taken as fact. In the absence of facts, as an indication of the right perception, we also lose any ability to justifiably castigate those whose expressions deviate from scientific laws.
Sentential Relations In Science
Logical Positivism, in part, is a response to needless groundless positing of unobservable events or entities, and assertion of meaningless claims. This movement was very focused on language and semantics, and conceived of the meaning of statements or terms as dependent on whether or not those statements can be tested. Carnap was quite critical of Heidegger (1959) in this regard, and often referred to his statement, "the nothing itself nothings," as an ideal example of the metaphysical tradition: assertions about the nature of reality that lack any sort of empirical verification. Positivists thought that neither metaphysics, nor most of philosophy, made claims of the testable sort. This led them to consider claims made by both camps to be meaningless.
The method that is applied to testing the meaningfulness of statements is referred to as the criterion of verification. Along with this criteria is the notion of two very important groups of terms and statements: observational and theoretical statements. Both the criterion and the statement groups contribute toward a view that aims to strengthen scientific claims by a logical analysis of language and observation. I will first discuss the criterion of verification and how it contributes to the positivist's notion of meaningful claims. Following this I will clarify the distinction between observational statements and theoretical ones, and how these groupings are thought to assist the criteria in enabling us to make scientific claims.
The Criterion of Verification
The criterion of verification asserts that the meaning of a sentence, if it is not logically true or false, consists in its capacity to be tested. As the Stanford Encyclopedia of Philosophy puts it, “The central idea behind verificationism is linking some sort of meaningfulness with (in principle) confirmation, at least for synthetic sentences.” Statements such as “the nothing itself nothings,” is meaningless on this account in that it makes no empirically discernible difference to physical facts, there are no physically confirming, or disconfirming, factors or conditions. Given that physical facts are key, positivists believe that experience and sense data are the sole means by which to assess scientific theories. Instances in which claims about gravity state it to occur will either be confirmed, or disconfirmed by observations we make. There is no way to specify what types of observations need be made, or will occur, concerning the nothing nothing-ing. It does not seem that any type of empirical observation will make any sort of difference in regards to Heidegger's statement. In the case of gravity, if we were to observe an apple fall from a tree, and then hover in mid air, the theory is greatly impacted. Also, the theory of gravity makes many predictions about what will occur to objects in open space above ground, all of which make some sort of difference about physical reality:

A proposition which is such that the world remains the same whether it be true or false simply says nothing about the world; it is empty and communicates nothing; I can give it no meaning.2

The tradition of metaphysics, primarily the dialog involved, is viewed by positivists as being saturated with untestable meaningless claims; this is not the same as being false. Consider a quote from Deepak Chopra,3 that “everything that we call matter comes from something that is not material.” He would continue to say that subjective states give rise to the physical universe. Given the context of the debate I believe what he means by subjective states is mental states. The two claims seem to compliment each other well, but how we would go about testing them is a conundrum. Having testable conditions, which often includes particular methods of measurement, is what enables a claim to be found either true or false, without which a claim is neither. From this positivists maintain that the claim is meaningless. How do you go about testing/measuring an incorporeal entity? Every observation we can make of physical entities makes no difference towards Chopra's hypothesized incorporeal source. There seems to be nothing by which to apply measurement or other means of physical evaluation about his claims, thus there is no means by which to verify whether his claims are true or false. This is what differentiates metaphysical claims from simply being false, and why positivists think it is erroneous to think of them as such: there is no means by which to evaluate whether such statements are true or false; even a false proposition is saying something about the world.
Keeping with the theory of gravity, the statement, “the apple that falls from the tree will drop to the ground (unless obstructed in some other physical way)” will either be observed as such, or it won't be. Either way, the claim will make some discernible difference as to how we understand the world: we know the conditions under which the statement will be true or false. These conditions will establish the meaning of the statement:

A statement is meaningful if and only if it can be proved true or false, at least in principle, by means of the experience. This assertion is called the Verifiability Principle. It follows that the meaning of a statement is its method of verification; that is, we know the meaning of a statement if we know the conditions under which the statement is true or false.4

Let's consider the claim “thing x is subject to gravity”. According to Carnap, the claim is meaningful if two conditions are satisfied.5 First, the mode of the occurrence of the subject, x, in its elementary sentence form (i.e. the simplest sentence form in which the subject is capable of occurring) must be fixed. Second, if the first condition is fulfilled, then it is necessary for it to answer the question:6 Under what conditions is the statement supposed to be true, and under what conditions false? We made mention above of one way in which it would be false, that being the apple hovers above ground and is not being suspended by anything. Likewise, there are conditions that would make the statement true. The initial statement “thing x is subject to gravity” meets Carnap's first condition, and is derivable from other statements of the same form: “x is a physical entity”, “x has weight”, and “x is contained within Earth's atmosphere”, to name a few. This is not to say that the statement is true, observation and measurement will be required and that's exactly to the point—there are stated conditions by which the initial hypothesis is true or false, all of which lend themselves to observation or other means of empirical evaluation. Any object contained within the Earth's atmosphere is said to have weight, and due to which, be attracted to other physical bodies. As such, if x were dropped from a staircase and not physically attracted to the ground by its weight, then we have a condition by which the initial statement is false (true in the case of a successful fall). The meaning of the word “gravity” is thus available, and determined by the answer to the question featured in Carnap's second condition.
Using verificationism as a foundational tool, positivists proceeded to develop a translation project that involved mapping statements to sets of confirming and discomfirming observations. The translation project resorts to linguistic analysis by appealing to an observational language that consists of terms for objects and properties detectable by the senses. It was already mentioned that positivists believe sense experience to be the only way to judge scientific claims, and say anything meaningful about the world. An issue for positivists is that there are terms and statements about certain objects and happenings that are not immediately available to the senses that seem vital to certain scientific enterprises: physics involves the study of nonobservable components such as energy and force. The terms 'energy' and 'force' aren't about things that can be observed, similar in that respect to metaphysical claims. At this point it would be best to advance onto the dichotomy positivists made between observational and theoretical statements, and how this distinction was meant to make the criteria of verification applicable to science.
Observational Language
As previously stated, the positivist movement was very focused on language and semantics, both of which serve as a means to make the world intelligible to ourselves. Scientific claims are meant to provide understanding about specific events and experiences, an understanding that is made available only through those statements that can be tested to be true or false. Chopra's notion that all matter originates from something incorporeal provides no such understanding. Claims like his do not lend themselves to testing, primarily because they do not have observable conditions. Positivists focused on claims that consisted of statements that are about observable occurrences that one may directly experience via the senses. Terms and statements of this sort were referred to as an observation language:

The positivists assumed (without much argument) that there is an isolable observation-language that refers exclusively to objects and properties that can be “directly perceived by the senses.” This vocabulary was essentially borrowed from everyday speech, and included such concepts as 'blue', 'hard', 'hot', and so on. At least some of the theorems derived from the axioms of a scientific theory must be couched exclusively in the observation vocabulary in order to be capable of empirical check...7

       The latter notion is essential to the positivist position in that verificationism is meant to provide us with theories that are empirically meaningful. Such terms as 'blue' and 'hot' are called observational terms in that they are words that denote objects or properties that can be directly experienced. In constructing an observational statement, one may include only observational terms. You won't ever find the term 'molecule' in an observational statement. “The stove is hot”, is an example of an observation statement in the that 'stove' and 'hot' are observational terms, each of which refer to things that can be directly observed (with caution). In the instance that an observation statement is true, we consider it a fact.
In addition to consisting of observational terms, observation statements can include logical terms, which contain all mathematical terms.8 Carnap recognized that the latter is not universally accepted as an observable, but he found it rational to consider inferred measurements as observable. An example would be to consider a given weight, let's say mine: 156 lbs (to be exact). He thought the philosopher would not consider this an observable property because the tradition does not consider it to be immediately available to sense perception. However, the physicist would include my weight as an observable seeing as it is a magnitude that can be observed through inference from simple techniques of measurement (placing me on a scale) as opposed to the mass of a molecule, the measurement of which is complex and indirect.9 Much like how I see a “red chair”, which denotes an observable object and property, “156 lbs” works in the same way in that simple weight measurements allow for a direct account of a things; “person x is Nemo” is derivable from “x weighs 150lbs”.
Empirical Laws
Consider the claim “All octopuses have eight arms”,  a statement that appears to pass the criteria of verification in that there are such entities that are octopuses, and they can be observed as having a certain number of arms. These kinds of statements are called empirical laws: simple generalizations we can directly confirm by observation. They explain facts and are employed to predict empirical facts by deducing observation statements from laws and statements of initial conditions.10 These laws are directly supported by observation reports, which are simply observation statements that pertain to a given law by being about germane observed events. Every individual account we have of octopuses provides an observation report that incorporates observation terms, such as “thing x is an octopus”, and “x has eight arms”. A law is formed from the generalization of collected observation reports that all describe the same happenings.
A problem for empirical laws is that no one observes all the octopuses in existence (both past and present), either all at once, or over a lifetime. 'All' operates as a universal quantifier, establishing a claim that is beyond verification given its inability to be evaluated via sense perception. One may consider to recover by adjusting the quantifier, that is, to speak of some, or at least one entity. However, this lacks the explanatory force that the universal has, which is the disposition science deals in. Empirical laws are crucial for scientific enterprises in that they aid the direction of certain studies, and allow for predictions and explanations of observation reports, providing understanding for specific facts and other directly observed events. To categorize laws as meaningless would be too damaging to science, however, Carnap (2009) thinks that we are misconceiving this situation: empirical laws are confirmed, rather than verified (p. 333):

At no time is it possible to arrive at complete verification of a law. In fact, we should not speak of “verification” at all—if by the word we mean a definitive establishment of truth—but only of confirmation.

His notion of confirmation accepts that we never observe a universal. Rather, we have a large number of individual observation reports that offer the same description of octopuses, and due to a lack of conflict, each offer support for the empirical law. This is not to abandon verification, it is still vital that an observation language be what it is because the statements involved can be tested. Carnap felt that the talk of verification gave out the wrong idea, and that no law is ever verified. A law acquires a stronger degree of confirmation with each instance of support, but we can never know for certain if future reports will coincide with past and current ones.
Theoretical Language
In my account of Carnap's inclusion of logical terms, I made quick mention of an entity that would be categorized as nonobservable: a molecule. Other entities such as electrons, atoms, protons, etc. are also nonobservable due to being neither directly perceivable nor available to simple means of measurement. Terms such as 'protons' and 'atoms' do not occur within the observation language, rather, they are theoretical terms that denote objects or properties we can only infer from direct observations.11 Despite not being available to direct observation, the methods by which we infer weight and temperature are extremely simple through appropriate devices: a balance scale and thermometer. The mass of a molecule is a different story, and requires complex means of measurement to the degree that it is considered indirect. Terms that would describe the mass of a molecule are not considered to be observational terms.
Positivists struggled not to lose theoretical terms to meaninglessness, and decided that theoretical statements had to include observational terms as a means of inference to conditions of verifiability. Theoretical statements include, but are not limited to, theoretical terms. Consider the statement “Water is H2O,” a statement that includes both an observational and theoretical term. This is considered different than “the nothing nothings”, in that “H2O” is placed in relation to a term that lends itself to observation, or direct measurements. There is much about this in positivist literature that is nebulous, but unlike observational statements, theoretical ones have to include other types of terms, observational and logical terms, but only by means of correspondence rules, which I will discuss in the following section. Other than these mixed statements, theoretical statements consist of theoretical terms, those terms that are set apart from observational terms by the verifiability principle.
Theoretical Laws
Carnap (2009) admits that there is no commonly accepted definition for theoretical laws, but his interpretation on the matter might be beneficial (p. 337):

A theoretical law is not to be distinguished from an empirical law by the fact that it is not well established, but by the fact that it contains terms of a different kind. The terms of a theoretical law do not refer to observables even when the physicist's wide meaning for what can be observed is adopted.

This is not to say that theoretical laws don't include observational terms, but primarily  they contain different types of terms. It seems Carnap would have to admit this in order for theoretical laws to not be completely meaningless. Similar to how empirical laws are vital to understanding facts, theoretical laws are used to gain a better knowledge of empirical laws. In the next section I discuss one version of the positivist notion of a scientific inferential relation between sentences. It involves an account of an explanatory relationship between empirical laws and observation reports. For instance, the law “all octopuses have eight arms” serves as an explanation for the individual reports “x is an octopus” and “x has eight arms”. But why do octopuses have eight arms? I assume we would have to resort to some form of gene/evolution theory for an answer, claims involving theoretical terms.
There is an initial dilemma in how theoretical laws are able to explain empirical laws. Despite having the problem of using universal quantifiers, empirical laws still consist of observable statements, and thus can make use of terms that meet the criterion. The terms contained in theoretical laws don't meet the criterion. By positivist accounts, the distinction between theoretical and observational terms depends on the criterion as a means to distinguish scientific claims from metaphysical ones. However, claims involving terms referring to nonobservables entities, such as 'molecules' and 'genes', are commonplace in the formulation of empirical laws. Theories involving genes explain why all octopuses have the observed shape and properties, much like how an empirical law explains our individual experiences of octopuses. Dismissing theoretical statements as meaningless would be quite damaging to scientific theorizing.
Comparable to how we confirm empirical laws, Carnap thought we can do the same, though indirectly, with theoretical laws. Empirical laws achieve confirmation by appealing to observable conditions. This is because these laws are about observable events and facts. Theoretical laws are not about observables, but about nonobservables. Recall that empirical laws arise from the generalization of facts, but seeing as theoretical laws are theoretical statements, and, as such, are not restricted to verifiable terms, they cannot be supported the same way. A theoretical law begins purely as a hypothesis, a supposition as the starting point for investigation. Carnap believes that from a hypothesis, we can derive a variety of hypothesized empirical laws, each of which can be tested by observation reports. The confirmation of those empirical laws derived from a hypothesis provides indirect confirmation of the hypothesis, which, in turn, becomes a theoretical law.
The empirical support for theoretical laws is considered indirect given that it relies on the confirmation of empirical laws, the latter serving as a channel to observation reports. Empirical laws need no such channel seeing as they consist of observational terms. While this may seem like a simple solution as to how confirmation reaches theoretical laws, another problem arises, that being the issue of how statements and claims from two different language types are supposed to be able to collaborate. Theoretical statements describe nonobservable entities and events, so even in the confirmation of a theoretical law, how are the terms contained within the theoretical language supposed to depict observable happenings?
Carnap, along with a few others, attempted to make functional a set of roles for the appropriate sentences to have that allow for empirical interpretations of theoretical statements. Sentences that work in this way are called correspondence rules, which mean to establish a correlation between real objects (or real processes) and the abstract concepts of a theory.12 Sentences of this sort are “mixed”, that is, they contained both theoretical terms and observational terms. The purpose of this is to [provide a link between theoretical and empirical strata of theory, allowing prediction and explanation in the downward direction (from theoretical laws, to empirical laws, to reports), and confirmation in the upward.]13 An example Carnap (2009) gives is “if there is an electromagnetic oscillation of a specified frequency, then there is a visible greenish-blue color of a certain hue” (p. 340). The statement itself serves as a rule that connects something observable with a nonobservable microprocess. The purpose of this rule is to provide empirical meaning to theoretical terms. The observable appearance of the green-blue indicates that a particular unobservable process is occurring. This is not to say that this specified electromagnetic frequency is defined by such observation terms, Carnap thinks this is not possible, to define theoretical terms with observable ones. Nevertheless, correspondence rules were thought to grant a solution to the problem of theoretical terms.
The Hypothetico-Deductive Account of Confirmation
Positivists saw scientific theory as a collection of sentences that stood in a certain relationship with each other, primarily one that is representative of symbolic logic. The principal format for scientific claims is a conditional, with the hypothesis considered as a universal condition statement: (x)(Px→Rx). Regardless of the type of format a scientific theory is organized into, one of the main positions of science is its ability to explain or predict phenomena. When the conditional format is put to work to predict or explain particular events the format is called the hypothetico-deductive account of confirmation. This specific format depends on at least one universal law from which a prediction is deduced. “All octopuses have eight arms,” is a law, serving as a hypothesis, from which we can predict that if we are told of an octopus in the next room, that it will have eight arms.
A hypothesis entails that particular confirming reports will occur: the observation statement about antecedent Px will be followed, or accompanied, by an observation report of the consequent Rx. If the report and initial observation statement/prediction match up, then the theory, or law, is confirmed. The epistemic import of observational evidence, then, is based on whether particular observation statements bear a certain logically formal relation to a law. The reason why the hypothesis initially operates as a law is that a countless number of individual observations have reported octopuses with eight arms:

If a certain regularity is observed at all times and all places, without exception, then the regularity is expressed in the form of a “universal law”.14

Carnap understood a universal law as being a collection of individual observations of a particular regularity, the reports of which describe the same conditions and properties of the regularity. In addition to this, there are no observation reports that describe the same regularity otherwise. A regularity of observation reports is simple enough to organize into a collection of support for empirical laws, but it is still rather unclear whether correspondence rules are sufficient enough to allow for reports to indirectly support theoretical laws in this way.
Carnap's version of the positivist translation project, hypothetico-deductivism (HD), is not concerned with whether a claim is true or false, though it's still crucial that it has the capacity to be, and he focuses on whether it has observation conditions, similar to the methodology of verificationism. Recall, however, that Carnap deviates from verification in regards to laws. He appeals to confirmation, his reasons for which come from the fact that regardless of observation statements within an HD structure relating to each other by necessity, thus are logically deductive, scientific practices are inductive by trade. Laws never observe universals given that no number of finite observations can guarantee the certainty of a law. The types of claims HD seeks to confirm are universal statements with the form (x)(Px→Rx). These claims are never proven true, seeing as the octopus in the next room could have six arms, and the same goes for the octopus in the room after that, and so on ad infinitum (though all you need is one disconfirming instance to undermine a law). Rather, they gain support from observation statements, or in other words, they help to confirm the universal.
In addition to being able to predict certain phenomena, Carnap also thinks that HD is explanatory. Sharing the exact same structure, explanations work like predictions in that both require reference to at least one universal law to help unite the given observation reports. The primary difference is that with predictions we are unsure if the consequent Rx will actually be observed after having observed the antecedent Px. With explanations, we already have both Px and Rx. What happens is that we come to understand Rx because of the law and Px is observed or occurs. Imagine someone who eats black grapes and gets sick. Unbeknownst to the victim of the nature of such a grape, he ask a doctor why he is sick, the doctor then responds, “You ate black grapes.” In combination with the law, “All black grapes will make you sick,” which can be symbolized as (x)(Bx→Sx), the occurrence of eating black grapes Bx, explains to the victim why he got sick Sx. Again, what establishes a law in the first place is a frequently reported regularity that (so far) has no deviating observation reports. Given that the law is formulated as a universal conditional, it means to obtain in any and all germane circumstances.
Bogen and Woodward's Response
The HD model of confirmation is conceived of, by positivists, to establish a two directional relationship between theories, or the statements that represent them, and observational evidence, described by observation reports. In one direction an observation statement representing a theory (Bogen and Woodward refer to such as “hypothesis statements”) entails that a particular experience will occur that is to provide a specific observation report. In the other direction an observation report, or series of which, describes events as they were experienced that confirm the hypothesis statement. The two directions amalgamate only two different categories involved in the scientific process: theory and observation. Bogen and Woodward15 criticize this view by claiming that a third category has been completely overlooked: phenomena, which stands in between the two. In addition to this, B&W do no think that the epistemic import of observational evidence has anything to do with inferential relations between sentences. I will start with B&W's castigating interpretation of the HD, and follow with their proposal for establishing a connection between theory, evidence, and phenomena.
Data vs. Observation Reports: What Counts as Evidence
B&W see the blueprint of HD to consist of theoretical claims represented by hypothesis sentences, and observational evidence represented by observation sentences or reports. These components fulfill first order logic requirements that allow for confirmation to take place on the part of observation reports in regards to hypothesis sentences. The conditional format that characterizes the HD demonstrates this in that a hypothesis statement, such as “all octopuses have eight arms,” which can be translated into the universal condition statement, (x)(Ox→Ex)16, entails that certain singular observational statements will come to be reported, in particular “a thing that is an octopus” and “it has eight arms”. If the observation report matches up with the observation statement that is predicted, after the appropriate experience occurs, then the report confirms the hypothesis statement. What B&W perceive to matter most in this scenario is that [the obtaining of the confirmation relation depends upon logical relations between]17 the various types of sentences involved.
        B&W consider HD to be a version of what they call “IRS”, or Inferential Relations between Sentences. The inferential relations that transpire in HD grant evidential relevance, in the form of confirmation, to observational sentences that fit properly into a first order logic format. As discussed in the prior section, in order for an experience to be regarded as evidence, it must be describable by an observation report (as such, it is the report that is the evidence), and the report must abide by the rules of first order logic. In contrast to this, B&W maintain that the epistemic import of evidence does not depend on an inferential relation between sentences, nor whether it is contained within an observation language. Simply because an observation report fulfills some aspect of a universal conditional statement does not entail that the observation report serves as evidence for a theory:

Just as entailment can hold between false as well as true sentences, confirmation can relate worthless evidence to unacceptable hypotheses as well as good evidence to correct or well justified theoretical claims. Just as the mere fact that p entails q does not tell us whether we should believe q, the mere fact that o stands in the required inferential relation to h does not tell us whether there is good reason to accept the claim h represents.18

The HD model does not seem to offer a means to determine whether an observation report should be accepted as evidence apart from meeting inferential requirements.
Carnap thinks that the law, “all octopuses have eight arms”, should entail particular confirming observation reports on the grounds that all previous reports have thus far supported the law. If we receive a contrasting observation report, then the law loses support, and perhaps is abandoned completely. You might think that the latter is what endangers the continual acceptance of a law, however, it is more of what counts as confirmation that enervates IRS models. In the case of (x)(Ox→Ex), the given law is logically equivalent to (x)(~Ex→~Ox): any instance of the universal confirms the universal, and given that the two are logically equivalent, to confirm one is to confirm the other. Under the parameters of equivalence, my blue cell phone helps to support that all octopuses have eight arms seeing as it is not an octopus and it doesn't have eight arms; it confirms (x)(~Ex→~Ox). On this account HD allows far too many items and observations to serve as evidence for a hypothesis.
It should be obvious that an observation report of my cell phone offers no insight into the theory pertaining to octopuses, and yet it fulfills first order logic requirements by means of logical equivalence, thus the report qualifies as being evidentially relevant. B&W stray away from the notion of observation reports as they think that not all evidence is represented in the linguistic manner that positivists conceive of, and the means by which reports come to count as evidence does not depend on sentential relations. Instead they consider data as evidence for phenomena. Data are [effects produced by elaborate causal processes that may involve the operation of the human perceptual and cognitive systems as well as measuring and recording devices and many other sorts of natural and manufactured non-human systems].19 Examples of this are the records of temperature readings used to chart various melting points, scores on psychology tests regarding memory processing, and photographs used to calculate the deflection of starlight. The same can be said for each of the examples, but to pick one, B&W do not consider a photograph to be a component of an observation language, on the contrary, it is a corporeal entity, and serves as evidence all the same.
Data is a crucial component in theory testing, and while it takes an evidential role in the process, it is not to be understood in the same way an observation report is. Not all instances of data pertain to an observation language: data are essentially not an observation report or sentence of any kind, they are objects in the world. As stated above, an example of data that is evidence for a phenomenon is an actual photograph, not some report of one. B&W think it indispensable for IRS methods to equate evidence with sentences to the extent that it is not an observation that supports a hypothesis, rather, it is an observation report. This results in excluding many vital pieces of evidence on the grounds that they have no sentential representation, that they are not contained in an observation language. Contrary to positivists' notions, the evidential value of data is 'extra-linguistic', and it is not necessary for data to have sentential representation to count as evidence.
Let's recall that, in the HD, theories not only predict observation reports, they also explain them, which characterizes its two directional format. B&W believe this is a flawed way to view actual scientific processes. It is not data, that theories seek to predict and explain, rather, it is phenomena. B&W interpret phenomena as objects, events, or processes in the world that yield naturally occurring stable patterns. Some examples are melting points of various substances, memory processing in the brain, and the deflection of starlight. Not exactly simple cases, but their complexity helps emphasize the character of phenomena. I found Glymour's (2000) interpretation to be quite helpful on this matter (p. 30):

The two differ epistemically because phenomena are, while data are not, explained by theories. Hence phenomena can be direct evidence for theories, while data, at most, can be only indirect evidence for theories. Further, knowledge of data is non-inferentially justified, while knowledge of phenomena is inferentially justified. To say that a scientist is wrong about the data she reports is necessarily to say that she did not in fact see what she claims to have seen, while to say that a scientist is wrong about the phenomena she reports need only be to say that she has drawn incorrect inferences from what she indisputably did see. The two differ ontologically in that phenomena are stable, repeatable features of the natural world, while data are not.

B&W believe positivists, such as Carnap, are mistaken in thinking that hypothesis sentences are in direct inferential relations with observation reports. Any theory, or law, does not explain any one single observation, which an observation report represents, rather, a theory looks to explain a pattern of occurrences. Glymour (2000) continues (p. 30):

The evidential relation between theory and data, insofar as there is one, is then composed of some inferential relation between data and phenomena (data are evidence for phenomena) and another between phenomena and theory (phenomena are evidence for theories).

The inferential relation mentioned by Glymour, between data and phenomena, is not an instance of IRS. On the contrary, data is in a causal relationship with phenomena. As mentioned, not all data can be represented sententially, regardless of which we are not looking for the relationship to be sentential when testing for the existence of phenomena.
The Epistemic Significance of Evidence
The types of elaborate causal processes that data are the effects of involve many parts and maneuvers that all play to various modes of simulation. These are not just simple acts or events of observation, though observation is a part of it, but they are the effects of an active engagement with various means of measurement and the environment (either immediate, or at great distance). When an observation report deviates from a law, B&W still don't believe this to be adequate grounds to dismiss the law because the issue is to determine the reliability of data based on the conditions by which they are produced. B&W (2003) consider the role the observation report plays in theory testing as extremely limited (p. 225): is often very hard to see how to construct a sentence which both represents the... evidence scientists use as data and also captures what is epistemically significant about them.

Consider that someone gives an observation report of a six-armed octopus. This conflicts with our original hypothesis statement, and thus the law loses its status. What of the person that gave the report, and the conditions under which they observed it? Was this person on any sort of medication that might distort their vision at the time? What type of vision do they have? Was the creature underwater (as water will distort vision)? Observation reports do not have the capacity to be epistemically significant in that positivists do not seem to take into account the manner in which such evidence is produced.
There are two features that B&W claim are required of data to make them epistemically significant. Firstly, they must possess features through which the phenomena in question can be studied, so my blue cell phone possesses no epistemic bearing in relation to the characteristics of octopuses. Secondly, data must have the capacity to be tested and analyzed in order to ensure they can inform us about a specific phenomenon. The way in which this is sorted out is by investigating the procedures and instruments that generate data. This is necessary to ensure that the data that are produced are causally related to the phenomenon being investigated. Establishing that there is a causal relation is what enables data to be informative about phenomena. If the investigation of the data generating procedures indicates little to no signs of error, then the data produced is considered epistemically significant. There is very little, if any, detail and consideration given in regards to how data is produced within the positivist outlook. They make no mention of such details, and can't be presumed to attend to offer an account of how to properly import epistemic data. This can't be presumed to be part of their framework given that, as B&W put it, not all the data can be represented sententially.
Upon considering Einstein's theory of general relativity, the use of photographs played an important part in determining the reliability of the theory to make predictions about the deflection of starlight due to the gravitational influence of the sun. The deflection of starlight is the phenomena under investigation, which means that the evidence that will be produced directly support it rather than Einstein's theory. B&W take into a account a specific investigation that sought to differentiate between three competing claims about the phenomenon: 1) there is no deflection; 2) deflection of the magnitude predicted by general relativity; 3) alternate measurement of deflection predicted by an augmented account of Soldner and Newtonian physics. Once the data that is produced establishes a causal connection with one of the three claims, then the outcome can be weight against Einstein's theory.
A couple of pair investigators, Curtis/Campbell and Eddington/Cottingham, are cited to have used photographs for data to test the three claims against each other. To ensure the dependability of the data, the data-generating procedures had to be inspected:

To interpret the photographs, investigators would have to establish their scale, i.e., the correspondence of radial distances between stars shown on an accurate star map to linear distances between star images on the photographs. They would have to measure differences between the positions of star images on the eclipse and the comparison photographs. They would have to calculate the deflection of starlight in seconds of arc from displacements of the star images together with the scale. At each step of the way they would have to correct for errors of different kinds from different sources.20

Complementing this are many different types of photographs taken to account for the possibility of errors: photographs taken at different points of the day and year, in various positions, and the proofing of star photos to establish scale. Each of these kinds of data require testing in themselves in order to establish whether or not they can be considered epistemically significant. This involves inspecting the cameras to confirm that they function properly. Even the adjustments made to the cameras in order to capture an image are surveyed: film type, aperture settings, shutter speed, etc. At one point certain complications arose with Curtis' data, the details of which we need not entertain for the point of this discussion. What is important is that it was determined that the procedures involved in producing Curtis' photographs were observed to contain errors. Due to this, the photos were considered epistemically negligible, they could not be depended on to be informative of the given phenomenon.
B&W maintain that the epistemic relevance of the photos is due neither to a mismatch between observational statements and observation reports, nor any sort of failed inferential relations between sentences. Rather, [the evidential value of the starlight data depended on non-logical, extra-linguistic relations between non-sentential features of photographs and causes which are not sentential structures.]21 As for the epistemic significance of the photos, the main requirement for data to have evidential value is for the detection process applied toward the data-generating procedure deems it as reliable. This generally means establishing a level of trust in importing phenomena claims as evidence for a theory by indicating the degree of support that data can offer phenomena. Trust is created by indicating whether the investigative device, or process, will provide data that will differentiate between phenomenon claims, and do so with a high rate of success. From this we're meant see how sentential representations are an 'unnecessary intermediary', and that processes focusing on raw data will produce better results in determining the epistemic value of evidence.
Nietzschian View on B&W: Nothing is Extra-Linguistic
Up to this point we have considered two philosophical theories concerning scientific practice. The first prescribes that observation reports, which are sentential items, operate as evidence for a theory, which is also sentential, by means of inferential relations between sentences (IRS). B&W reject this method, claiming that theories comment on phenomena, not data, and the latter support phenomena by a causal relationship that has nothing to do with IRS. Both positions maintain that a good deal can transpire in science with or without language, and in each case we think to establish an epistemic relationship between us and our environment.
In this section I present my interpretation of Nietzsche, from which I argue that both of the above positions are untenable. I draw mostly from “On Truth and Lying in an Extra-Moral Sense” (1873), in which I understand Nietzsche to think that language is what the human intellect assimilates, not pure unvarnished things apart from human interests:

The “thing-in-itself” (which would be pure, disinterested truth) is also incomprehensible to the creator of language and not worth seeking. He designates only the relations of things to men, and to express these relations he use the boldest of metaphors!22

Human comprehension is nothing other than finding ourselves reflected in things. The relations of things, which is formulated in a way so as to cater to our intellect, is done with language. As such, our formulations do not report on actual physical affairs, instead they reflect our interests.
       A point I wish to make in this section is that we mistake expressions we have habitually designated for things as descriptions rather than socially accepted metaphors. Given this, in science, epistemic claims are apocryphal as it concerns stating the reality of things. I will elucidate on how Nietzsche's ideas undermines B&W's conception of how we understand data, as well as how this hurts the IRS. Beginning with B&W, I will offer my understanding of what it means for data to be extra-linguistic. Then I will discuss how language is needed to assimilate data, within this I clarify Nietzsche's idea of the peace agreement that we appeal to when making epistemic claims. Proceeding from this is a discussion of how when language is employed towards “the world”, we are doing nothing other than creating relationships between certain terms and statements, and that language does not extend beyond itself.
Following my criticism of B&W, I will proceed to discuss what I think is troubling about positivism from a Nietzschian perspective. Some of the problems are shared with B&W, but while B&W believe to work without language, positivism incorporates language in far too serious manner: language is not literal, rather it is metaphoric. As a result of this, I present Nietzsche's brief view on how we inverse time so as to make it seem concept terms, such as “heat”, are the cause of certain impressions, when it is really the result of a creative impulse. Along the way I will make use of Wes Alwan's interpretation of “On Truth and Lie” to assist my notions.
What is it for data to be 'extra-linguistic'?
Recall that data are the effects of elaborate causal processes. One example B&W give are the photographs produced in the analysis of Einstein's experiment. We can imagine various types of expressions to demystify data, but no such expression would be an instance of data seeing as it's an “object in the world”, not a term or expression. I interpret B&W to mean that data are things-in-themselves, objects that are not related to human cognitive capacities and interests. There are two conditions by which data are extra-linguistic:

1. Data achieve their evidential status apart from language, or IRS: this is troublesome because if some feature of data is to serve as evidence, then it must be a feature that we can refer to.
2. Data are not linguistic items in nature: Given this, what is data? It cannot be “photograph” seeing as data is not linguistic; any term meant to pertain to it is leading us astray. The dilemma here is a matter of admitting the need for language, but then whether correspondence is had between terms and things.23

       Though data is only one aspect of their theory, it is vital to understanding phenomena. Unless B&W want to admit that data is epistemically lost to us, they will have to allow that a socially fixed language is needed in order to reference aspects of data that are to be evidentially relevant. However, even in doing so, statements elaborate only on other statements, or singular terms; language does not extend beyond itself to things-in-themselves. If language is an indication of our interests, then it seems absurd to think it can correspond to and represent something it is not: some disinterested thing.
There Are No Intelligible Extra-linguistic Objects
Pertaining to the first condition, by what other means do we make things intelligible to ourselves than by utterances such as “hard”, “cold”, or “photograph”? Nietzsche (1989) thinks that the only things that humans assimilate are those things that are relatable, that are human-like (p. 251-252):

Just as the astrologer observes the stars in the service of men and in connection with their joys and sorrows, so such an investigator observes the whole world as linked with man; as the infinitely refracted echo of a primeval sound, man; as the reproduction and copy of an archetype, man. His procedure is to hold man up as the measure of all things...

If data really are extra-linguistic, that is, they are not the result of our interests, of having ourselves as the measure of things, then we have no means to make them intelligible to us. Thus far our discussion of data has shed light on it with language, but seeing as data is not linguistic, it seems like any term meant to pertain to it is leading us astray.
For Nietzsche, the reality of our affairs is nothing other than narration, or the transfer one's own desires or emotions—language is the creative construction of our affairs. To investigate the way the world works, is simply to be engaged in language:

...where there is chattering, there the world lies before me like a garden. How lovely it is that there are words and sounds! Are not words and sounds rainbows and illusive bridges between things which are eternally apart... how should there be any outside-myself? There is no outside. But all sounds make us forget this; how lovely it is that we forget. Have not names and sounds been given to things that man might find things refreshing?24

Our chattering is the human landscape; words and statements are the reality insofar as they are the means to create coherence between the multitude of our various impressions. While I doubt B&W will concede that data are nothing other than linguistic items, it seems they will have to admit that language, and perhaps a system of established sentential relationships, is necessary in order for us to comprehend the data.
The peace agreement. Whatever data may be, for us, cannot be extra-linguistic in that we desire to intellectually possess it. For that to happen there must be conventionally fixed terms and phrases that have been appropriated to fit seemingly similar experiences so as to give us a sense of organization concerning our environment:

… because man, out of necessity and boredom, wants to live socially in the herd, he needs a peace agreement... this peaceful agreement apparently leads to the first step towards man's acquisition of his mysterious desire for truth. For what “truth” will be from now on is fixed; a uniformly valid and binding terminology for things is invented and the legislation of language also enacts the first laws of truth.25

Associating “stove” with “hot” depicts not truth-in-itself, but signifies the success of the peace agreement in that we are reiterating the associations we have been taught to make. If an infant were given free range on the sounds she wished to make concerning her impression, who knows what kind of reality she would construct. However, to ensure that the intellect has something to digest, we each have to be made to make certain utterances towards seemingly similar impressions, so as to make it seem we understand those impressions. To have knowledge of our environment is not discover that a “stove” is “hot”, but it is to do well in playing copy-cat with utterances made in the past. To “interact” with our environment in regards to the comment of another, such as if someone were to ask you to bring him water, and you comply in the manner expected, is not to demonstrate that language communicates understanding concerning the desires of others and the objects of our environment. All this demonstrates is compliance with the peace agreement, that you have been well trained in associating certain words correctly; you know how to play the game as it is expected of you.
In the same scenario, if you were to bring him something other than what has been socially fixed as “water”, he will feel disgruntled, and since he is playing by the rules, he will have the social token of justification on his side in upbraiding you. Perhaps you brought him what is customarily a “book”, to which you say, “This is water!” Nietzsche thinks that there are two different lies transpiring here. The first is to equate terms with things perpetrated by the conformist. For the sake of sustaining the peace agreement, this lie is necessary and unconsciously accepted. The second is the association of “water” with something that is conventionally fixed as “book”. This is the unforgivable lie in that it violates the social norms that bring us comfort. The latter is the type of deception that produces harm or inconvenience:

The liar uses the valid terms, the words, to make the unreal appear real; for instance, he says, “I am rich,” when “poor” would be the right term. He misuses established conventions by arbitrary substitutions and even reversals of names. When he does this in a selfish and damaging manner, society will no longer trust him and so it will exclude him from its presence. But men flee not so much being deceived as being harmed by deceit.26

It cannot be that the conformist fears deceit, for he is the benefactor of it. The comfort that comes from this deceit can only be had by unconsciously accepting the peace agreement, to act as though terms correspond with things. “Water is H20” is the kind of deceit that benefits the social player in that it provides the intellect with a sense of assimilation towards things.
While language may be necessary in grasping data, Nietzsche (1989) expresses great doubt in whether language is capable of getting to reality at its joints (p. 248):

...what is the situation with those conventions of language? Are they perhaps products of knowledge, of the sense for truth? Do terms coincide with things? Is language the adequate expression of all realities?...

It should be obvious that he does not think it is. If we read off the structure of reality directly from it, unmediated, cognitively, and turned that into language, then there wouldn't be so many different languages. Nevertheless, language is what we have to work with, and in order for data to be something the intellect can work with, it must be linguistically represented. In this sense, data is not extra-linguistic.
Words Do Not Refer or Correspond to Things
B&W may not deny that words are needed in order to make data intelligible for us, but words are not what we are working with when it comes to observing a causal relationship between data and phenomena. Concerning the second condition of data being extra-linguistic, “photograph” is not what we're really dealing with, but it is the thing itself. However, language is an interest-laden tool that science is attempting to grasp disinterested truths with. Language cannot be in relation to something apart from human interests given how utterly opposite it is to the objects we mean to reference:

When we speak of trees, colors, snow, and flowers, we believe we know something about the things themselves, although what we have are just metaphors of things, which do not correspond at all to the original entities.27

Words like “impressions” and “reactions” are not things we have in regards to an impersonal world, instead they are artful utterances that we use, unbeknownst to the devout follower of the peace agreement, to relate to ourselves—when the scientist investigates she only finds an extension of herself in properties such as “cold” and “hard”:

If someone hides an object behind a bush, then seeks and finds it there, that seeking and finding is not very laudable: but that is the way it is with seeking and finding of “truth” within the rational sphere. If I define the mammal and then after examining a camel declare, “See, a mammal,” a truth is brought to light, but it is of limited value. I mean, it is anthropomorphic through and through and contains not a single point that would be “true in itself,” real, and universally valid, apart from man. The investigator into such truths is basically seeking just the metamorphosis of the world into man...28

Our devotion to the peace agreement has led us to think we discover “water is H20”, as though the statement would hold in our absence. Nietzsche does not think this is the case, and instead we designate things as such, and then forget this so it seems it is something we have found. As such, concepts appear to us as the result of things we have discovered about the world.
We believe we become epistemically familiar with things by using agreed upon terms, such as “leaf”, to reference an object, which seems to establish a correspondence between language and things based on habitual use. The repetition of a term is what forms a concept:

Every word becomes a concept as soon as it is supposed to serve not merely as a reminder of the unique, absolutely individualized original experience, to which it owes its origin, but at the same time to fit countless, more or less similar cases, which, strictly speaking, are never identical, and hence absolutely dissimilar. Just as no leaf is ever exactly the same as any other, certainly the concept “leaf” is formed by arbitrarily dropping those individual differences... this gives rise to the idea that besides leaves there is in nature such thing as the “leaf,”...29

Concepts are a means to take the things of our individual experiences and universalize them; to think of a group of impressions as similar so as to abstractly equate them. One might be tempted to think that when uttering “leaves” in his quote, Nietzsche himself is appealing to language in a way that fits particular instances of experience, that he is using this term because it picks out specific physical things. I don't believe him to take this association seriously, and is simply appealing to the peace agreement so as to be able to converse with the conformist. To utter “fitting” or “corresponding” in regards to an impression is to appeal not to our “knowledge” of how things happen, or a “fact” of the matter, but is an individual interpretation using the predetermined concepts that make up the peace agreement; a relationship is created between “term”, “fit” and “individual cases” constructing a picture that is completely dissimilar from whatever it was that we experienced.
The word, “photograph”, is conventionally associated with seemingly similar impressions only by overlooking the differences between the impressions. One may claim that we make sense of data, despite it not being "photograph", because our familiarity with the concept allows for us to reference and pick out the impression in question. This, in-itself, is considered to indicate a correspondence between terms and things. There are two problems with this. Firstly, our familiarity with the concept by no means grants us an understanding of the object the term means to correspond to. To be extra-linguistic is to not be a word or sentence in any way. So in saying that we are familiar with the concept, we're claiming to be familiar with what the object is not. As for our ability to pick out or refer to a particular object, this itself says nothing about the object itself, nor does it indicate an understanding of it, rather it is indicative of how we are taking our interest-driven associations as something other than metaphoric assignments. This tends to happen due to the repeated use of concepts, but Nietzsche (1989) does not think compliance with the repeated use of a concept epistemically justifies it (p. 252-253):

… when the same image has been produced millions of times and has been passed down through many generations of men, indeed ultimately appearing to all mankind as the result of the same occasion, in the end it has for man the same significance as it were the only necessary image and as if that relationship of the original nerve stimulus to the produced image were a strictly causal relationship—just as a dream, eternally repeated, absolutely would be felt and judged as reality. But the hardening and solidification of a metaphor is not at all a guarantee of the necessity and exclusive justification of this metaphor.

Our chatter, our transference of interests, construct our reality. Why should instances of interaction between ourselves and “things” be examples of linguistic correspondence? Of course the explanation will be through terms whose very correspondence is in question. “Interact”, itself a concept meant to refer to seemingly similar occurrences, does nothing other than correspond with other terms. In this case it's sandwiched between “humans” and “things”, two terms that are thought to relate with physical things outside of themselves due to our pretensions towards the concept “interaction”. Due to equating things that are dissimilar, we are not actually acquiring knowledge of things. We might think B&W evade such deception by arguing for objects in scientific processes that aren't contaminated by human interests. Nonetheless, Nietzsche (1989) thinks that deception is the only game in town when it comes to making things intelligible, and the intellect prospers by forgetting this (p. 252):

Only by forgetting that primitive metaphor-world... only by the invincible belief that this sun, this window, this table is a truth-in-itself, in short, only insofar as man forgets himself as a subject, indeed as an artistically creative subject, does he live with some calm, security, and consistency.

Even if B&W admit that concept words are necessary to make sense of data, concepts do not refer to things-in-themselves. Alwan's (2014) interpretation of Nietzsche clarifies this (p. 83-85):

… all that is “given” is the chaotic world of images to which we have to give order by stamping ourselves on it – as “artistically creating” subjects – by means of concepts... the idea of immediate access to things-in-themselves – or correspondence between our representations and things-in-themselves – cannot be made sense of. The only possible relationship we could have to things-in-themselves is “aesthetic,” in the sense of our constructive activity in giving concepts to impressions.

The very tool that is necessary to the intellect, language, is unable to reference things beyond what it has been used to construct.
Nietzschian View on Positivism: Taking Language Too Seriously
Positivist notions concerning the role language has, in epistemic pursuits, especially the criterion of verifiability, been quite influential. While positivism recognizes the necessary role of language, in this section I will elaborate on how Nietzsche thinks positivists are mistaken in thinking language to function literally in regards to things outside of itself. Nietzsche offers a brief view on the inversion of time that I interpret as a reason why we are taken in by language in a way to think that concepts cause impressions, rather than being the result of a creative process. I will explain how the contents of the observational language only receive membership by positivists taking statements literally—“the stove is hot,” counts as observational insofar as it's accompanied by two beliefs: that words correspond to things outside of themselves, and that words and statements precede impressions instead of being produced by them. In addition to this, as discussed above, we form concepts of things by neglecting the unique aspects of individual experiences, and make it so that the words resulting from an individual experience correspond to countless diverse experiences.
All Language is Metaphor
Nietzsche (1989) views language to be the result of sudden sense impressions (p. 248-249):

First, he translates a nerve stimulus into an image! That is the first metaphor. Then, the image must be reshaped into a sound! The second metaphor... the origin of language is not a logical process, and the whole material in and with which the man of truth, the scientist, the philosopher, works and builds, stems, if not from never-never land, in any case not from the essence of things.

Each stage serves as a metaphor for whatever it is we are exposed to, because neither translation is the same as what was experienced. While Lakoff thinks that [our ordinary conceptual system, in terms of which we both think and act, is fundamentally metaphorical in nature] (Wikipedia 2014), this seems to coincide with Nietzsche's ideas. However, the two diverge on at least one point, that being Lakoff's notion of non-metaphorical thought. For Nietzsche there is no non-metaphorical thought; where ever there is language, there is metaphor. Enervating our notion of reference is not to doubt our impressions, rather it is to mistrust knowledge, and everything that is designed to comfort the intellect. The peace agreement, and the concepts contained therein, originates from our drive to survive, and is mistaken as being an indication of epistemic progress.
For positivists, the meaningfulness of statements is how we make the world intelligible to ourselves. Observational terms such as “stove,” and “hot,” are thought to denote objects or properties that can be directly observed. When we have particular sensations, positivists appeal to a group of terms that have been fixed given past experiences, and select the germane terms to be the cause for our sensations. If I experience a sharp pain in my hand, “I touched a stove, and it's hot,” I appeal to the terms that the criterion of verification denotes as being available to my type of experience, the observational language. Like all matters, I'm not appealing to a disinterested world. The intellect is provided with an understanding of things through the observational language in believing the statements contained therein are the cause of my impressions rather than being the result of the two-stage metaphoric process.
We are twice removed from our initial impressions. Each stage is in no way similar to the properties of the things that inspired the impression. Observation reports do not describe our experiences. Instead they are personal interpretations, a subjective composition of metaphors (concepts) that indicate a creative relation between sentences, not an inferential one. As such, the deductive framework of the HD is an arbitrary construction of statements that appears logical insofar as they have been rehearsed enough to be accepted:

… what would allow us to say, “The stone is hard,” as if “hard” were known to us otherwise than as a subjective stimulation! We arrange things by genders, we designate the tree [der Baum] as masculine, the plant [die Pflanze] as feminine: what arbitrary transferences! How far-flung beyond the canon of certitude!30

Observational terms are fixed insofar as they are the result of some past metaphor that we've come to reiterate towards things and properties that seem similar. This is successful only by overlooking nerve stimuli of the moment, the translation of which may not match the terms used in the past. It's easy to imagine that “atom” could have been an observational term if it had been the intuitive translation of some past impression, and then reiterated under similar circumstances. These reiterative measures are the foundation of making the environment intelligible.
Cause and Consequence Confused
Nietzsche's translation project interprets that human expression provides for itself an object of knowledge; the “intellect” as a concoction of what absorbs such object. It does this is by organizing our metaphors in a manner that makes it seem as though they are embedded in the things of our impressions:

… on to a certain sensation, the result for example of a distant cannon-shot, a cause is subsequently foisted (often a whole little novel in which precisely the dreamer is the chief character). The sensation, meanwhile, continues to persist, as a kind of resonance: it waits, as it were, until the cause-creating drive permits it to step into the foreground – now no longer a chance occurrence but as 'meaning'. The cannon-shot enters in a causal way, in an apparent inversion of time. That which comes later, the motivation, is experienced first... The ideas engendered by a certain condition have been misunderstood as the cause of that condition.31

We believe “stove” and “hot” are the literal translation of our experience, as the cause of our sensations. This engenders the idea that statements precede nerve stimuli rather than being their result. Observational terms, having been the metaphor of some past experience, are memorized and retrofitted to a new experience, and rather than accept that our utterances are the result of current nerve stimuli; “hot” is thought to be what we experience that causes certain impressions. The moment we are impacted by an impression, the two-stage process quickly results in our sputtering out, often times, words. We no longer allow our instincts to create our expressions, which potentially [brings about new metaphors that undermine the existing system of concepts, always threatening to refashion it into something like the more chaotic and colorful world we see in dreams (Alwan 2014, 105-106).] Instead, the intellect recalls some past seemingly similar stimulus, in regards to a current impression, of which we have been trained to associated a particular concept to it. Rather than accepting no two impressions match,32 and due to which, concepts have not a foot to stand on, we craft a story of affairs to make it seem as though “hot” precedes the impression to make it seem like we assimilate the cause of it.
Our drive for truth, for scientific security, does not allow for new experiences. Nietzsche does not think we have the courage to accept what we might intuitively translate with each experience. It's far safer to translate through memory, [to trace something unknown back to something known is alleviating, soothing, gratifying and gives moreover a feeling of power.]33 What is known is simply a reiterated metaphor. We forget that our statements are the twice removed consequence of a past experiences, not the cause or explanation of a current one; they are simply recycled utterances that cater to the desires of the intellect.
The predictive feature of HD, then, is not to foretell, or explain, new experiences, but to rehash past metaphors. “Octopus”, or any other concept, appears as an object of knowledge, but is nothing more than a banal metaphor of which we forgot that it had been the twice translated consequence of some other experience. Observation reports do not give fresh accounts of the current experiences, for if they were to, no two reports would ever match. Inverting time, and the peace agreement, both aid in this inferential deception. The intellect is driven to organize the world in a manner that is reasonable, where effects are preceded by causes, and in the case of positivism, where statements fit like pegs into the formulations of first order logic to produce laws that explain our experiences. In any case, we have equated things that are completely dissimilar: metaphors with things in the world, and our individual experiences.
Language has not only been used to create an object of knowledge, but a whole system of “truths” by which we organize our lives by. We work in metaphors of which we have forgotten they are metaphors. Forgetting is our salvation in by doing so we can revel in the “intellect”, and everything it can possess. The two scientific theories I discussed present two different ways in which we can be objective about the world. Observational language establishes a set of terms that do more than simply express our interests, they are terms that we can empirically test, thus they describe what we experience. Orchestrating language in this way helps us feel epistemically disposed toward claims that are meaningful due to their relevance toward physical affairs, while all other assertions are frivolous and negligible. If we want to be artful in what we say, as so many science fiction authors are prone to do, we must accept that our prose is not serious in relation to observational statements, as though the artist's spirit doesn't inhabit the tongues of scientists as well. I believe this dichotomy of terms, of our human landscape, is what fuels the fires of castigation and ridicule, of punishment; the child can utter the “sky is octopus” for only so long before it is demanded of her to be serious.
Observation reports are facts that add up to tell us why it is, and what will be. In this respect science leeches off of the past, off of the peace agreement, so as to avoid making new metaphors, new art. Someone who is not disposed toward their creative instincts finds sanctuary in one of two linguistically rigidifying practices: science or religion. Nietzsche thinks both to be ascetic in that both repress their abilities to create worlds with words. Instead they are each in the practice of reiteration, of appealing to the peace agreement so as to set their “truths” apart from the ones initially created for survival purposes. As Nietzsche (1967) says of the scientist (p. 148-150):

“These Nay-sayers and outsiders of today who are unconditional on one point—their insistence on intellectual cleanliness... these last idealists of knowledge in whom alone the intellectual conscience dwells and is incarnate today---they certainly believe they are as completely liberated from the ascetic ideal as possible, these “free, very free spirits”; and yet to disclose to them what they themselves cannot see—for they are too close to themselves... They are far from being free spirits: for they still have faith in truth.

Ironically, B&W dedicate just as much faith in the literal function of language in that the things science works with are actual “photographs” and “objects”. Similar to positivists, our words are thought to correspond with impressions in a way that depicts how the world really is. The very things that are not terms are thought to be depicted by terms. Metaphors dance only with other metaphors, and the extra-linguistic object is a mystery to us. Even in its “natural” form it is of no use to the intellect seeing as it cannot assimilate it apart from our interest-laden faculties.
Whether it be a series of observation reports that confirm a law, or extra-linguistic data that support phenomena claims, science attempts to give an account of the world that is free of human bias. This is an absurd endeavor seeing as “objectivity” itself is born from out interests, our need to understand our impression in an unequivocal manner. The objective view, the one that is able to step outside the human psyche to “see things for how they really are,” is itself nothing other than a linguistic construct. At the end of it all, science is still children at play, “the sky is blue,” or “is purple,”, except that these kids, along with the rest of us who take language literally, are taking the game too seriously. This is not to say that Nietzsche is opposed to survival, as language has enabled us to do, nor is he opposed to deception, as this is what the human landscape is founded on. There is much beauty to be had in what we construct, and the process of doing it can be invigorating. What he is critical of is the jump that we make from survival-driven expressions to an epistemic depiction of life. Both create a system of social punishment, only the latter is absurdly comforted by intellectual justification.

1. James Bogen and James Woodward, “Evading the IRS,” Poznan Studies in the Philosophy of the Sciences and the Humanities 20 (2003): 236.
2. Moritz Schlick, “Positivism and Realism.” In Logical Positivism, ed. A.J. Ayer, et al. (The Free Press, 1959), 76.
3. “The Future of God Debate,”
4. “Mauro Murzi's pages on Philosophy of Science,”
5. “Mauro Murzi's pages on Philosophy of Science”
6. Or any equivalent formulation of the same question.
7. Timothy McGrew, trans., Philosophy of Science: An Historical Anthology (Blackwell Publishing, 2009), 316.
8. Mauro Murzi's pages on Philosophy of Science”
9. Rudolph Carnap, “Theory and Observation.” In Philosophy of Science: An Historical Anthology, ed. Timothy McGrew et al. (Blackwell Publishing, 2009), 337.
10. “Internet Encyclopedia of Philosophy,”
11. “Mauro Murzi's pages on Philosophy of Science”
12. “Internet Encyclopedia of Philosophy”
13. Timothy McGrew, Philosophy of Science, 321.
14. Carnap, “Theory and Observation,” 329.
15. From this point forward to be referred to as “B&W”.
16. Where “O” stands for “is an octopus”, and “E” for “has eight arms”.
17. Bogen and Woodward, “Evading the IRS,” 228.
18. Bogen and Woodward, “Evading the IRS,” 229. (“o” refers to an observation sentence and “h” to a hypothesis sentence)
19. Bogen and Woodward, “Evading the IRS,” 224.
20. Bogen and Woodward, “Evading the IRS,” 232.
21. Bogen and Woodward, “Evading the IRS,” 234.
22. Friedrich Nietzsche, “On Truth and Lying in an Extra-Moral Sense,” in Friedrich Nietzsche on Rhetoric and Language, ed. Sander Gilman, Carol Blaire, David Parent (New York: Oxford University Press, 1989), 248.
23. This latter point is a problem for positivism as well.
24. Friedrich Nietzsche, “Thus Spoke Zarathustra,” in The Portable Nietzsche, ed. Walter Kaufmann et al. (New York: Penguin Group, 1976), 329.
25. Nietzsche, “Truth and Lying,” 247.
26. Nietzsche, “Truth and Lying,” 248.
27. Nietzsche, “Truth and Lying,” 249.
28. Nietzsche, “Truth and Lying,” 251.
29. Nietzsche, “Truth and Lying,” 249.
30. Nietzsche, “Truth and Lying,” 248.
31. Friedrich Nietzsche, “Twilight of Idols,” in Twilight of Idols and The Anti-Christ, ed. R.J. Hollindale et al. (London: Penguin Group, 1990), 61.
32. This is not to say with certainty that experiences can't be identical, but how are we to know when our means of making things intelligible is the metaphoric consequence of nerve stimuli?
33. Nietzsche, “Twilight of Idols,” 62.


Alwan, Wes. A Guide to Nietzsche's On Truth and Lie in the Extra-Moral Sense, Kindle Edition,          2014.

Bogen, James, and Woodward, James. Evading the IRS. Poznan Studies in the Philosophy of the Sciences and the Humanities 20, 2003.

Carnap, Rudolph. “Theory and Observation.” In Philosophy of Science: An Historical Anthology, ed. Timothy McGrew et al. Blackwell Publishing, 2009.

Glymour, Bruce. “Data and Phenomena: A Distinction Reconsidered,” Erkenntnis 52, 2000.

Internet Encyclopedia of Philosophy. Logical Positivism. (accessed Dec. 11, 2014).

McGrew, Timothy, trans., Philosophy of Science: An Historical Anthology. Blackwell Publishing, 2009.

Murzi, Mauro. Mauro Murzi's pages on Philosophy of Science, (accessed Dec. 11, 2014).

Nietzsche, Friedrich. “On the Genealogy of Morals,” in On the Genealogy of Morals and Ecce    Homo, ed. Walter Kaufmann et al. New York: Vintage Books, 1967.

Nietzsche, Friedrich. “On Truth and Lying in an Extra-Moral Sense,” in Friedrich Nietzsche on Rhetoric and Language, ed. Sander Gilman, Carol Blaire, David Parent. New York: Oxford University Press, 1989.

Nietzsche,Friedrich. “Thus Spoke Zarathustra,” in The Portable Nietzsche, ed. Walter Kaufmann et al. New York: Penguin Group, 1976.

Nietzsche, Friedrich. “Twilight of Idols,” in Twilight of Idols and The Anti-Christ, ed. R.J. Hollindale et al. London: Penguin Group, 1990.

Schlick, Moritz. “Positivism and Realism.” In Logical Positivism, ed. A.J. Ayer, et al. The Free Press, 1959.