Friday, December 10, 2010

Trust has no place in science #2

Thanks to all who replied and shared their views. Particular thanx to Christina who replied in her blog. With Saml, and Cameron and Bill they think this is about semantics. Linguistic tricks. I hope not; this is too serious to get away with such. "Reliable, trustworthy, assumptions": it's all working around the real issue. Similarly, splitting up 'trust' into 'blind trust' and 'smart trust' is just working around the real problem.

Indeed, my point is different. The key of science is to replace trust by facts. Or, when talking about database, software, research papers in Nature, it is replacing trust with traceability. Actually, we seem to have lost a long-standing tradition of citing previous work when we write down the arguments we base our argumentation on. Facts are backed up with references, providing the required traceability.

Now, compare that to current electronic sciences. We 'trust' our database to have done something sane. Well, don't. They made an attempt, but made errors. As they say with software, having zero bugs just means you have not found them yet.

The real point with 'trust' is, is that it is completely irrelevant. It adds zero to the scholarly discussion. Whether you trust the highly curated ChEMBL database or not, it has errors. (Noel pointed out one source of ambiguity in the ChEMBL database this week). What does matter, instead, is if those errors are significant. Do they affect the conclusions I draw when I use this data. That is what actually matter. Trust has no place in science. Error has.

Sadly, this is basically the hypothesis of the VR grant I wrote up but did not get awarded. But I trust I do better next time.

Why this matters? Well, this is what ODOSOS is about: bring back the traceability into science, and get rid of trust.


  1. I think it is naive to believe that trust can be totally removed from science. One simply does not have time to check every detail and to follow every assertion back to its roots. Indeed, I am not sure one could do this completely - because there are so many dependencies.

    I think the key comes from a quote from President Regan, "Trust, but verify." This is where ODOSOS come in - it gives one more opportunity for verification. Following the ideas of the Reproducible Research movement and having everything in a compendium makes this easier.

    A scientist who uses ODOSOS makes it easier to earn my trust - in other words, when they make it easy to check their work and what I check out is well done and makes sense, I can decide whether the time required to dig deeper is profitable or whether I trust what I don't check out on the basis of what I did investigate.

  2. "I think it is naive to believe that trust can be totally removed from science."

    But is it relevant that you trust it or not? You use it or not, that's the only thing that matters. Not your (or anyone's) trust in it.

    Removing trust from science does not mean you must verify everything. Those things are not each others opposite.

    "One simply does not have time to check every detail and to follow every assertion back to its roots."

    And there is no need for that. You can use whatever shoulders you like to stand on. Whether you trust those shoulders or not, that's irrelevant; if you trust some shoulder, that does not make it better or worse.

    Likewise, getting back on trust and verification not being opposites, verifying something does no change the correctness of something either!

    Look at it like this: do you trust gravity? I doubt you do. You just accept it. It has never failed you. What does trust add here?

    To counter Reagan, I would say "Use, and, if you want or need to, verify".

    To me ODOSOS is not about adding trust at all. ODOSOS is about me being able to see where sources of error are, how large those errors are, and what I can possible do to change that. Perhaps even see what I can do to enlarge the application domain of some ODOSOS component. That is why I care about ODOSOS.

  3. There is perhaps another way I can try to express myself: 'Trust' is emotional instead of objective. Trust makes you feel good. But science is not about feeling good. It is about facts. Feeling good about gravity is nice, but adds nothing.

    Someone feeling he must check every fact, likewise, is emotion too. Checking, however, is not an emotion, it is an action taken after feeling bad or suspicious.

    But, there are other reasons to check facts too, that do have a place in science. For example, on reason to check your input data, or facts your reused, is that the conflict with new results.

    I see no need to use 'trust' (or similar) here in such scholarly activity. It only adds emotion to something that should be done objectively. So, I uphold that trust has no place in science.

  4. I like what you said "Use, and, if you want or need to, verify." Practically speaking, that is what we have to do.

    But I dislike the assertion that there are "facts" in science. I prefer the word "data" because it implies that what we know is only what we have measured so far.

  5. Hi Ms.PhD, you are absolutely right! Jean-Claude made me aware of my flaw in arguments just a bit earlier:

    As I replied there too: facts are emotionally tainted too. It does not matter if something is a fact or not, it is the observations that are used as arguments that matter.

  6. You might consider Executable Paper Grand Challenge

    It's all about automatic verification and reproducibility of published results.

  7. I commented elsewhere, but let me reiterate. I work in biology. I need to trust the chemistry literature/databases to 'get it right.' Only if my work has failed spectacularly - and repeatedly, because I distrust 'new' work first - would I go back and try to get someone else to confirm the structure of something like a small molecule drug already FDA approved. If I can't access a database I trust which will hand me that structure, the work will stop because I don't have the facilities or skills to go sort out that chemistry on my own. Same thing, let me add, for atomic weights and physical constants. I'll doubt my measurements of bat wings before I verify g, for example.

    So, it is important that the chemists come together and sort things out. Improve or shut down specialty databases. Go do some editing on wikipedia for structures with public interest. The biologists need to do similar things with quite a few of our own resources; we can be trustworthy.