As a follow-up to my last post, it looks like I should offer a more detailed explanation of why exactly scientific activity is a group activity -- not simply as a matter of convenience, but as a matter of necessity. Helen E. Longino has already made this case very persuasively in her book Science as Social Knowledge (specifically the chapter called "Values and Objectivity"), so I'm going to use this post to give a sketch of her argument.
The upshot of the argument is that objective knowledge requires the involvement of other people in the building. All by yourself, there is no way to move beyond subjective knowledge.
First, what do we mean by "objective"?
Subjective has to do with how things seem to me. Objective, in contrast, has to do with how things actually are - what the world is really like regardless of how it might seem to me.
There's another sense of objective that we think is linked to this first sense: what is objective is what anyone could see. In particular, it's what anyone could see using the appropriate method for getting rid of the subjective stuff. The appropriate method for seeing the world objectively, we might imagine, is simply the scientific method.
We already have a general sense of this methodology: you design an experiment, calibrate the instruments, make accurate measurements that your labmates check, then use these to build or assess your hypotheses and theories. While some of the pictures of science (like Karl Popper's) allow as how certain moments in science might be more subjective - say, in the imaginative task of coming up with new hypotheses or theories - they assure us that the end product, scientific knowledge, will be objective because of the objectivity of the testing. You determine what the theory predicts, you set up an experiment, and you determine whether or not the predicted outcome actually occurs. By applying the rules of science to the data, the scientist is supposed to be able to get an accurate picture of what's really going on.
Kuhn, among others, has argued that the real workings of science are not this straightforward. Different scientists, applying the same rules to the same data, can get different pictures of what's going on. Plus, it's hard to completely dispel Kuhn's worry that observations are theory-laden. This means that my experiments don't give me facts anyone could observe. Rather, they yield facts that reflect how the world seems to me through the lens of my theory.
Is objectivity out of the question, then? Longino doesn't think so. She says objectivity is where the community structure of science becomes very important.
All by myself, all I can get is how the world seems to me.
In a group, I can find out what others see when they look at the same parts of the world. I can tell them what I see. Not only can we compare the features we agree and disagree on, but we can also go back and look again at the same parts of the world to see if they look any different now that we know what others see there. Peer review is a mechanism within science where this kind of communication takes place.
What's the point of the peer review process? The goal is to figure out what we agree upon and to filter out the influence of subjective preference as much as possible. Which parts of how the world seems to me are due to the world, and which are due to my subjective preferences? The parts we tend to agree on might be the best candidates for the features of our experience that correspond to real features of the world.
The theory-ladenness claim is overblown. Yes there are theories that observations rely on, most of the time. But they are rarely the theories that are being observed for, so to speak.
There is a general problem that arose in the context of evolutionary epistemology - there is no principled reason to think that a community will be any more objective in its views than a single individual, if the utility functions of the communal knowledge base happen not to coincide with the truth functions of the domain. In other words, natural selection, whether of ideas or of genes, tracks fitness, not truth.
So a communal standard can still be merely an internal realism, and not an external one. Peer review filters unfit ideas - those that fails to satisfy the utility functions of that discipline. It's a large leap from that to scientific realism.
As an outsider, it seems to me that subjectivity is an extraneous variable that peer-review corrects for in order to arrive at objectivity. Right?
It's true, as John says, that a community standard isn't necessarily any less biased than an individual's standard. But scientists use it for two reasons. First, on average, it is less biased. Second, it's the best standard we have.
I'm not sure on the theory-ladenness claim, but I agree with John regarding the internal vs. external realism claims. Since peer review functions within the community, it eliminates the individual subjectivity you describe. Unfortunately, that's all it does-- it doesn't assure objectivity.
On the other hand, I do think that it's just about the best we can get. I don't see a method for producing knowledge that's any more reliable and practical than the practice of scientists.
In a related question, what do you think the minimum size of a community of researchers has to be to ensure the sort of objectivity we can achieve?
As suggested by John, intersubjective agreement isn't sufficient to get us to objective knowledge; and I doubt very much that it's necessary. We could, of course, just stipulate that only the ideas that win wide acceptance in a discipline count as scientific, but then we'll have to whigishly write the many false leads and failed experiemnts out of the history of science. Isn't it easier to accept that an individual scientist might come to possess an item of knowledge before persuading her peers that this is so?
Wilkins's point that intersubjective agreement is not sufficient to establish that what you agree on is how things really are is well taken.
However, if you're going to claim that intersubjective agreement is not necessary -- not for happening onto how things really are, but for having a good basis for believing you've got a handle on how things really are -- I think you need to show me a positive argument for that.
In any event, the scientific commitment to avoid self-deception would seem to make the involvement of others (with both sense organs and rational powers) a very important thing.
As I read her, Longino isn't trying to come up with a way to get or guarantee objectivity in the first sense ("what the world is really like"). She's too explicitly following in the footsteps of Habermas, Foucault, Harding, Mill, and Peirce to have external realism or truth-as-correspondence as a goal. Instead, I think, she's starting with the assumption that subjective preferences (read: value judgments) are (a) held irrationally, and therefore both (c) not subject to rational debate and (c) liable to cause distortions and disagreement when they are used to support factual judgments (such as those science is normally supposed to be after) -- they lead to dogmatism.
Then her argument is that, because of (c), we need a way to filter out the dogmatic effect of value judgments. But by (b), we can't have a rational debate about them. So, instead, we'll get everyone looking over our shoulders, to check and make sure we haven't brought value judgments into play when producing factual judgments. That is, our subjective preferences haven't been able to distort our science. The argument is essentially Mill's of section 2 of On liberty, but the conclusion is Habermasian -- what the peer review processes produces is science that is acceptable to everyone, whatever their value judgments. So the goal is not truth; it's epistemic acceptability. (When she adds something truth-like to the picture, in her later book The fate of knowledge, it's also not objective in the first sense, but instead a very pragmatic property of "conforming to its object sufficient for the community's goals" -- paraphrase from memory, not a quotation.)
In short: If you want science that corresponds to what the world is really like, Longino's procedure is clearly irrelevant. But if you think value judgments are epistemically problematic, at least when they show up in science, she offers (as far as I know) the only currently viable account of how to filter them out.
As an outsider, it seems to me that subjectivity is an extraneous variable that peer-review corrects for in order to arrive at objectivity. Right?
One would hope so, but a manuscript in peer review will typically only be assessed by 3 or 4 people (an editor and 2 or three referees). Of course they are subjective too, and the referees can often come to wildly different conclusions. It should reduce the subjectivity, but perhaps as much by acting as a check to filter out subjective statements by insisting on applying accepted standards (e.g. speculation is frowned upon, and should be backed up by evidence).
I suspect the shift of a knowledge claim from being subjective to objective occurs after publication, as it gets picked up and mutated by the larger scientific community: the individual subjective assessments get averaged out simply because a large number of people are involved. I think this is more or less what Noumena is describing.
I agree with the first comment from John - that community is no guarantee of objectivity. Good examples of this could be the ongoing argument about the competing low-fat and low-carb dogmas or the many cases of supposed drug efficacy based on dubious interpretation of studies. My point is that communities, just like people, are subject to external influences and as individuals differ in the extent of their integrity. Some people are influenced by vanity or pride, others by pecuniary reward. Whatever the reason, the truth is just as malleable when the influences are strong enough and these influences can apply to whole groups as well as to individuals.
Trust me, I'm not knocking intersubjective agreement - it is by far the best approach we have ever had as a species and the success of science licenses a small amount of inductive confidence. I was merely pointing out that you don't get "objective" by adding together many "subjectives".
Janet - I think it's perfectly reasonable to hold that Newton knew things about the composition of light before he reported his discoveries to the broader world, and that his knowledge was scientific.
i envy the quality of your commenters, Janet, and thank you for an interesting post. i think several of the above replies skirt around some more basic issues, which are, what are we trying to do here? what do we hope to get from scientific inquiry? what is its purpose? that is to say, rather than take the fact of objectivity as a given and then ponder how ones gets it, we might consider the role and value of the concept itself. when doing so, the intersubjective part (which makes Harding more Habermasian than she is) and the theory-laden part are still important but not as primary questions. (and actually, i shouldn't say skirt around, since Noumena offers an admirable sally in this direction.)
for my own sake, i'll offer this set of posts which take a different view of the objectivity question.
in the least, i hope you can get Daston and Galison's (2007) Objectivity into this thread.
ugh -- i wrote harding, obviously meant to say longino. i'll let someone else deconstruct that slip.
Between Janet and Noumena's posts, I think that we've arrived at a consensus. Peer review in science keeps the scientific community epistemologically responsible, but it doesn't guarantee truth. As truth seems to be something that we don't seem to be able to guarantee, epistemic responsibility looks like it's the best we can get.
Did I miss something, or is that it?
I think you've changed your definition of science from the last post to this one. Here you seem to be defining science as the generation of objective knowledge about how the world works. In the prior post the definition seemed to be theorizing, observing, and testing hunches with the goal of gaining a clearing understanding of how a piece of the world works.
These are very similar but this posts definition requires teamwork and the last post doesn't really.
I think part of the difference is generating knowledge that is useful outside of how or why it was generated in the first place. (It might even be fair to claim that the former is scientific but isn't science and define the term as the later.)
In reality these are almost two different steps. You spend time in the lab doing an experiment--step one. Then you spend time either writing a paper or in someway packaging and explaining what and why you did for other people--step two. I think the definition from the earlier post allows step one to be science while the definition from this post requires both steps.
Flat Earthers complain that measurements, etc. by Spherical Earthers are theory laden.
In 1978 we revised Austrofundulus, and recognized two species. One had several kinda distinct local populations which we discussed but did not think merited taxonomic recognition. In 2005, we revised Austrofundulus, with DNA information added in, and recognized a total of seven species, including all but one of the 'kinda distinct local populations'. It amuses me how different those populations look now that I know they are separate species.
I'm a little worried that intersubjectivity as a feature of an epistemic process is being conflated with intersubjectivity as a feature of an epistemic object in some of the discussion here. It's true that adding together a lot of 'subjectives' doesn't give you an 'objective'; but when we're talking about intersubjectivity, that's like saying that a bunch of visual systems don't add up to a visible object -- true, but it tells you very little. Likewise, it's true that communal processes don't guarantee truth or objectivity (of what is arrived at), but it takes a rather serious skepticism to demand that they do so, since pretty much nothing provides such a guarantee.
I think Janet's original suggestion that objectivity (of the process, i.e, seeing the world objectively) is a sort of modalized intersubjectivity (of the process, ("what anyone could see using the appropriate method for getting rid of the subjective stuff") has a bit more bite than it might seem, because if we reject it, it becomes rather difficult to see what distinction is being made at all when we distinguish objectivity (of process) from subjectivity (of process). And given this, it's not such a large leap from intersubjective processes to realism -- a leap, yes, since it involves an assumption or two -- it's just basic causal inference, i.e., it's just a question of what the best causal explanation is for conclusion X having survived a process of eliminating merely subjective factors as its cause, and then it's entirely plausible to say, as Janet says, that features flagged by this objective fit to evidence are our best candidates for the real features of the world, on causal principles. That just requires us to assume that there are real causes conforming to very basic causal principles.
I wrote about these issues back in December.
One can also consider the question of what "community" comprises. Granted that any given manuscript might be prepared after consultation with only a very limited number of scientists, and then that manuscript might be peer reviewed, formally, by only a very few more, there is then the fact that "public"ation of the manuscript has made public its author's claims. The manuscript and the proposed knowledge therein then becomes the object of scrutiny by a much larger community, and, in the present day, this means that the scientific knowledge proposed must stand the test of scrutiny of scientists (and others) from a wide range of cultures, which means that the whole group is now almost certainly NOT subject to the same set of biases, pressures, desires, or utility functions.
Over time, the knowledge proposed either succumbs to falsification, or it does not. I don't know of any true scientist who claims to have absolute knowledge; the scientific vision with which I am familiar suggests only that what we call scientific knowledge is that which has not yet been falsified, and the longer the scrutiny goes on without falsification occurring, the more confident we are in the validity of the knowledge. Sometimes, of course, small inconsistencies or objections start to pile up, and when enough problems are identified, old "knowledge" gets replaced by a new explanation fitting the facts. A paradigm shift occurs, and the process starts all over again.
Community, then, when it comes to de facto peer review of any scientific finding, consists not only of all those people who interest themselves in or rely on any particular scientific finding in the era contemporary to that finding, but also of all those who come afterward and until such time as the knowledge is falsified. Community is not, then, a matter of two or three scientists sharing a lab or officially peer reviewing a manuscript, but rather it is a matter of an expansive number of people spread not only over place, but also over time.
> I was merely pointing out that you don't get
> "objective" by adding together many "subjectives".
I think this is both true and misleading.
You don't get truly "objective" by adding together many "subjectives" that operate off of the same set of preconceived notions. However, there are two additional factors involved here.
The first is that the preconceived notions can be limited and still provide a functional model. Newtonian mechanics are demonstrably an incomplete model of mechanics, as they break down when you get really fast, really small, or really massive. However, within some reasonable definition of "useful", Newtonian mechanics are functionally correct. While a bunch of physicists who assume subjectively that Newtonian mechanics are correct will have a hard time moving their paradigm to include quantum mechanics, that doesn't mean that their subjective evaluation is *incorrect*, it's just *incomplete*.
Second, very few explorations into science can be contained ultimately within a particular field. You might be a biologist, but you're going to be doing chemistry at least by proxy, or you're going to be relying upon subjective "truths" accepted by the chemistry field as part of the body of knowledge you use to advance your theory. Since, ultimately, all science overlaps into some other field somewhere, you're going to be coming across bits of information that can be verified using subjective "truths" accepted by some other body of science. Similarly, you can find a result in one field that relies upon some assumption that is commonly not considered to be part of objective truth in another field - biology contributes to the understanding of chemistry, chemistry contributes to the understanding of physics, and so on.
In a meta sense, however, the whole discussion is muddled by the concept of "truth". You can't really have "truth" in science. In order for you to have objective truth in the Q.E.D. sense, you have to have declared axioms and logic... that's mathematics, not science.
In my opinion, science isn't about a search for "truth", it's about exposing new phenomena to observation and building models that are "usefully correct", as opposed to "objectively universally correct". Given that you're dealing with (objects and speeds and distances) within this set of (possible objects, speeds and distances), they will react with some delta of (Newtonian mechanics) with each other. If they're outside of the set of possible ranges, this model is incomplete, go do research and build a new model that describes that set of ranges.
But that doesn't imply that the old model is *wrong*, except in the sense that some people (and this includes, unfortunately, some scientists) regard the model as objectively true. That's a problem with people, not science.
From this perspective, "subjective" and "objective" aren't really the right sets of terminology to use. "Subjective" carries with it a connotation of "is based upon opinion", and scientific theories aren't based on opinion, they're based upon observation, corroboration, and useful models. "Objective" carries with it a connotation of "is universally always correct", and that simply is never going to apply to science, because you can't build the Ultimate Theory of Everything.
More on the , and the concept of truth.