Please enable JavaScript.

Coggle requires JavaScript to display documents.

Objectivity in Science - Coggle Diagram

- - - - (1) Separate Virtues Problem (i.e. that TRUST and RULES are separate virtues in the process)
        
        By examining processes, we believe that there are certain rules in place to allow researchers to produce an objective result
        
        But does having rules mean that we should have trust in the product? The link isn't clear (example of errant marking in MCQ)
        
        Also consider Kuhn's SSR
        
        Kuhn suggests that scientific revolutions happen when we get a change in paradigms, and he rebuts people's claims that science is being driven by irrational factors by asserting that this process of change is governed by rules
        
        But, there's really no clear-cut way of saying 1 paradigm is better than another, even if we had a rule-governed process!
      - (2) Conflicting Virtues Problem
        
        What do we do when it ends up that the processes have led to an errant product/outcome?
        
        Should we just stick to the old outcome, because it has already generated trust anyway?
        
        In policy making, we often desire processes that regulates interactions FAIRLY; so we are okay with having rules that don't tell us how the world really is
        
        Case Study of NICE:
        
        When it was proven that their rule of not spending > GBP30,000/QALY for drug funding was wrong (the figure should be GBP13,000), NICE stuck to their guns and ignored it
        
        Argument that it is easier and more convenient for all to just stick to something we've already known/used for a long time (we have trust there already)!
- - - - The trick:
        
        The value-free ideal specifies that the internal reasoning practices of science must be value-free
        
        So, when you have evidence/data, and you are trying to draw out conclusions, NEVs cannot interfere with this 'core of reasoning'
        
        Here comes the problem of INDUCTIVE RISK
        
        Rudner 1953: Because science works by induction, claims will always be underdetermined (i.e. evidence will never give 100% certainty to a claim) --> Scientists must risk accepting false hypotheses, or rejecting hypotheses that are true, and they appeal to NEVs to make the trade-off
        
        Douglas 2000: Specifies that there is UNCERTAINTY in the collection of evidence and this causes you to need NEVs to determine what mistakes matter, and allow you to better obtain a resolution (so it occurs throughout the process)
        
        Case Study: Betz on the IPCC
        
        The big question is whether the IPCC and the report it produces are ways of proving that scientists need not take inductive risks by using HEDGED hypotheses, and qualifying their degree of uncertainty
        
        Concerns for Betz (See L2 John)
        
        2nd Order Uncertainty: Claims about how uncertain you are can matter
        
        Deferred IR: Seems like it just becomes a case where the scientists defers the taking of inductive risks to the policymakers
        
        Evidence: The IPCC's choice of WHAT evidence to include seems to include the use of NEVs (they ignore things simply because they were not peer-reviewed)
        
        L2 Betz
        
        Argues that the VFI can hold up in science as long as scientists make their uncertainties explicit by using hedged hypotheses
        
        Betz seems to think that scientists are required to produce plain hypotheses in order to produce policy-relevant science
        
        But, like in IPCC reports, scientists can communicate their uncertainty to avoid taking IR
        
        On the 2nd order uncertainty: Betz believes that this isn't policy-relevant (we take many empirical statements as plain facts without questioning their validity because we see their policy relevance as they are)!
        
        Personal Thoughts
        
        The process of articulating the uncertainty in the report seems to require the use of NEVs (what makes something 60% uncertain, or 50% uncertain)?
        
        Betz is focussing solely on the part about statistical significance, but seems to neglect the other internal processes of science (evidence gathering, interpretation)
        
        And I question the policy-relevance of hedged hypotheses
        
        L2 John
        
        Argues that the IPCC case study does not show that scientists can do science while adhering to the VFI
        
        Betz gives a false binary that if you use plain hypotheses, you must use NEVs (and vice-versa)
        --> Even with hedged hypotheses, they only are value-free IF, the hypotheses, relative to the evidence claims, does not involve significant errors/problems
        
        John notes that for the IPCC (Steele's argument), scientists still need to translate uncertainties to a qualitative measure
        
        Also, being a consensus-making body, the IPCC needs to have a framework for secreting evidence (which involves NEVs)!
        
        Finally, asking if a value-free IPCC report is valuable to policymakers (is policy impotence a price worth paying for value-free science?)
        
        Important
        
        In order to undermine Betz, the thing is to show that climate scientists MUST take inductive risks, that this is something they cannot AVOID! (rather than simply showing that this is something that they DO)
        
        One way to consider this is to use Steele's argument that the scientist QUA policy advisor takes inductive risk!
        
        L2 Winsberg
        
        This argues that uncertainty quantification (UQ) in climate modelling doesn't really achieve its goal of separating epistemic from normative values
        
        In climate modelling, the systems are so complex and intertwined that it is impossible to avoid calling climate modellers to take inductive risks --> Because they need to interpret the data, choose what model to use etc!
        
        Trying to screen these values simply doesn't work because of the sheer complexity of these models, and so, in short, one can never avoid using NEVs in the process of generating evidence for climate science! It is very much value-laden
        
        L2 Havstad and Brown
        
        Uses the case study of the IPCC's Pragmatic Enlightened Model to show how scientists can take inductive risks and STILL be objective/policy-neutral
        
        The PEM model acts like a policy map, whereby scientists list out all of the viable policy pathways based on their different predictions for policymakers
        
        IT explicitly admits that science is value-laden by proposing policies, and yet, leaves the decision-making to policymakers
        
        But, its issue is that this map ends up being very complicated, and may lose its value. It's also not an attempt to solve the VFI (not that it matters)!
        
        pp. 110 also specifically explains why climate scientists have a responsibility to make value judgements
        
        But, note also that PEM is merely just Betz's argument in a different form (it also calls for the deferral of value-laden decisions to policymakers)
        
        Appeal to the Authority of a Collective Community
        
        Oreskes further adds that science and knowledge is the consensus of a well-established community that produces these knowledge through a process of organised scepticism
        
        Thus, when we say science is an appeal to authority, it is an appeal to the authority of a collective community
        
        This can make it such that values and objectivity becomes compatible
        
        L2 Douglas
        
        Argues that the presence of AIR means that scientists must employ NEVs when doing science, particularly when the science done has political or social repercussions
        
        NEVs are important when scientists are trying to recognise whether it is better to have a false negative or false positive
        
        Here, we focus on the role of NEVs on the internal reasoning processes of science: (1) choosing levels of statistical significance (2) characterisation of evidence that is borderline (3) interpreting results for social outcomes!
        
        L2 Rudner
        
        Rudner makes a distinction between using values in the INTERNAL and EXTERNAL processes of science, and for him, he centres on the extrapolation of data to conclusions
        
        Underdetermination means scientists always ask: "how sure do we need to be before we accept a hypothesis" --> This depends on "how serious the mistake is"
        
        Fact-value dichotomy seems to be untenable (as per Quine), because the scientist qua scientist often makes value judgements about the facts they have
        
        What is needed instead is for scientists to be explicit about the value judgements they are making!
        
        L2 Steele
        
        Exapnsion on Rudner's argument by arguing that Rudner neglects how scientists as POLICY ADVISORS need to consider various practical implications of their suggestions to decide how wide their credibility gap should be
        
        As policy advisors, there are new dimensions of the scientist having to be strategic, and this strengthens the arguments that scientists need to use NEVs! --> Strengthens Rudner, because now, it's not just about the truth of the world, but also about the POLICY consequences of science
        
        IPCC: Scientists need to use NEVs to convert their beliefs from probabilities to something usable
        
        Reporting: Scientists need NEVs to choose how they should report intervals, because they need to recognise the pragmatic consequences of being wrong
        
        Regardless, someone MUST transverse the fact-value line; it boils down to an issue of division of labour!
- - - - Contextual values: NOT epistemic commitments because they aren't meant to be true! They are just... ways of looking at the world
        
        The contextual values are the key to understanding the importance of going social!
        
        Because these can be problematic, they need to be criticised, assessed and discussed as community norms
        
        Science is objective insofar as these contextual values are made explicit, discussed openly!
        
        Longino's 4 Conditions for Producing Good Contextual Values
        
        Venue for open criticism
        
        Uptake of criticism
        
        Public standards for assessment of evidence and of criticism
        
        Tempered Equality of Intellectual Standards
        
        Longino is calling for science to be more diverse, but for her, she argues that there is an epistemological benefit to diversity because we can refine our contextual values!
        
        The whole point of going social is to ensure that the contextual values at play are reliable/accurate
    - - Mainly, there are 2 types of biases at work: adaptationist and androcentrist
        
        Longino's account helps us understand this case study because here we are also talking about contextual values, and promoting the incorporation of feminist epistemologies into this field to strengthen/correct these contextual values
        
        L3 Llyod Ch 8
        
        Here, Llyod highlights the 8 assumptions that are evident based on the theories of female orgasm. Most of them are about the relationship between female and male sexuality (that the male body is a model), or the connections between orgasm and sexual intercourse
        
        This relates to 4 contextual assumptions/values, but Llyod focusses on adaptationism and androcentrism to explain why the byproduct model is often ignored
        
        Llyod uses Longino's 4 criteria for good contextual values, and argues that in the case of the female orgasm, criteria (3) about standardised knowledge/accounts had been violated
        --> Because we have the ardent adaptationist views vs the byproduct view, which adopt very different standards of evidnece/evaluation!
        
        NB: Note that Llyod is accused of being anti-feminist for promoting the byproduct view, but it can still be seen in a feminist light (there is legitimate feminist values in disentangling women/their bodily functions from their reproductive function)
      - L3 Llyod INTRO
        
        Here, Llyod is explaining the typical processes through which evolutionists try to prove that something is an adaptation for human/species survival
        
        But Llyod also notes that other than this adaptation view, some biological processes can be explained using a by-product model (e.g. the male nipple)
        
        For the female orgasm, Llyod argues that the accounts provided by scientists don't hold up to available evidence, or that they seek to reduce orgasm to merely orgasm with intercourse
        
        This case study will show the presence of biases in science, and recognise the importance of having alternative epistemologies to point out these biases
        
        It matters, because if we stick to an adaptationist view when its not true, women who don't have an orgasm during sexual intercourse would be viewed as dysfunctional... even though they are normal
    - - Perhaps it is more a matter of getting a normative account of how good values can be generated, which can be more reliably expected to produce good science.
        
        No account can give us a surefire way to GOOD science?
    - - L3 Longino Ch 6 (Socialising Knowledge)
        
        Here, Longino moves to show how that knowledge that a community produced (Ch 5) can become legitimate/epistemically acceptable!)
        
        Longino explains that underdetermination produces a gap that scientists solve through using contextual values, but these values too suffer from underdetermination
        
        This doesn't mean that those values are bad; as long as there are social process guarding the applications of these assumptions, science can still be objective (4 social norms)
        
        So, diversity has BOTH a moral and epistemic value
- - - - How it links to objectivity:
        
        A claim is STRONGLY objective when: it is well-established from every perspective possible
        
        A claim is weakly objective when: it is correct from within a singular standpoint
        --> Many scholars support this notion that objectivity is achieved when people from different perspectives agree on something! And this is not IMPOSSIBLE to achieve (it's hard, but not impossible)
        
        Alexandrova on Strong Objectivity:
        
        Alexandrova argues that some claims are mixed claims, in that they are causal, but have moral/political dimensions to them
        
        In such cases, the value claims must be made explicit, and then checked with people affected by the work to see if it is something they endorse. (so, the claim can be objective even if there are values, as long as the concepts are agreed from different perspectives)
        
        Nozik's INVARIANCE to explain strong objectivity:
        
        Given some fact/claim, if it remains invariant under different forms of admissible transformations, the fact/claim can be said to be objective
        EG: The claim that WWII ended in 1945 is strongly objective, because no matter how much we debate about what caused the end of WWII, we all agree that it ended in 1945
        
        So, there is a sense that objectivity is when perspectives CONVERGE
        
        Concerns:
        
        How do we decide WHOSE knowledge/perspectives do we consider? (e.g. conspiracy theories?)
        
        It seems then that agreement itself doesn't give us truth, and is more than mere agreement.
        
        L4 Nozik
        
        Here, Nozik describes what makes a fact or belief objective, and the significance this has on science
        
        For him, there are 3 conditions that make a fact objective: accessible from different angles, intersubjective agreement and indepdence
        
        So, with the concept of invariance, a fact is objective IF, under admissible transformations, it remains the same
        
        Importantly, for Nozik, subjectivity-objectivity is a spectrum, not a binary: we can speak of DEGREES of objectivity
        
        Objective beliefs are not assessed the same way as facts; a belief is objective when biases that leads it further from the truth are absent
        --> Important: Not about being FREE of value; but just having good values to counter biases to affirm objectivity
        
        For science, it is precisely that observations are value-laden, or they are path-dependent, that allows it to progress! (it is a cumulative effort after all)
        
        L4 Harding Ch 1
        
        Harding is tracing postwar global changes and developments to argue that changing political environments generated new philosophies of science --> That in a climate with social democratic movements, theories/ideas of PoS shifted too (new perspectives were created)
        
        An exceptionalism stance of science emerged post-war, as S/T became critical to the global Cold War, and scientists wanted to assert the autonomy of science
        
        1970s: Failure of development projects, counterculture (anti-authoritarian social movements), globalisation!
        
        In this volatile climate, new, alternative epistemologies emerged: feminist studies, postcolonial studies, SSK! They sought to show how science and their societies co-produced each other
        --> Harding asserts that these 3 case studies have shown how diversity strengthens objectivity, allowing us to achieve strong objectivity
      - NB: Harding and Longino
        
        Longino focusses on the social processes, trying to explain HOW social processes can lead us to objectivity
        
        Harding isn't interested in the interactions of different standpoints, but instead, is asserting that for claims to be objective, they need to have this strong notion
    - - L4 Wynne
        
        Argues that scholarly literature has neglected the intellectual value of lay knowledge. We cannot ignore its importance if we are to understand how natural knowledge is produced from social contexts!
        
        Rebuts Gidden's argument that a lack of public dissent means public trust in experts/systems! (it could be an issue of transparency, or technological fatalism on the public's view)
        
        Risk is a social construct manufactured by the expert institutions, and so, risk, to the public, is about the behaviour and trustworthiness of experts!
        
        We must stop seeing the lay-expert divide as a divide, so as to redistribute power AWAY from expert institutions
        
        The case study of the Cumbrian sheep farmers show how scientists ignore local knowledge at their own peril, as this increased distrust in scientists. Also, farmers followed the scientists' advice not because they trusted them, but they had no choice. And also, the scientists' failure to respect the farmers led to the questionable implantation of certain scientific policies or experiments!
        
        Case study of Andean potato farmers also shows the strong contrast between local, indigenous knowledge and universal scientific knowledge. Van Der Pleog explains that indigenous knowledge is often ignored, because it is seen as incompatible with standardised scientific knowledge!
        
        Link to P6 Science and Activism
        
        Why does indigenous knowledge matter, or what is the role of the public in a risk society?
        
        Also consider how combining science with local/indigenous knowledge can make it easier for sicentists/policymakers to accept it!