Jump to content

Talk:Predictive power

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

About the solar eclipse in 1919: I've heard that the results of the measurements were faked? I'm not questioning general relativity at all, but that this actual "confirmation" was not really true.Unclevortex 12:21, Jun 4, 2005 (UTC)

Harry Collins & Trevor Pinch's The Golem (ISBN 0521477360) provide an account of the experiment; the results were not faked as such, but the results of starlight during the eclipse, recorded on photographic plates, were highly variable and inaccurate, and could also have been interpreted as being a confirmation of the Newtonian model (which also predicted some bending of light too, but to a lesser degree). The Royal Society (or some similar body) arbitrated - some results were discounted, and others were included, with some quite curious logic used to apply as to which ones should be used; they decided that Einstein's propositions held. In science studies the case is often held up as a classic example, not of predictive power, but of how results can be selectively interpreted to confirm existing prejudices. In fact this article could not have used a worse example!
All the above is off the top of my head, I don't have the relevant text and sources to hand; when I do (in a few days) I will update the article. Qwghlm 23:51, Jun 5, 2005 (UTC)
I've added a disputed template to the page because of the above; anyone willing to add some more historical context to the general relativity example is welcome to - I would myself but am unable to do so right now. Qwghlm 00:06, Jun 6, 2005 (UTC)
Thank you for clearing that up! That's what I've read before too. --Unclevortex 10:17, 6 Jun 2005 (UTC)

Eclipse

[edit]

I'm well aware that the historical case of the 1919 eclipse observations is very complex (I'm a PhD student in the history of science myself), but as a brief example it works well and is one I think people will be fairly familiar with. I'd been wanting to write a full article on the eclipse experiment one of these days, because it is really quite fascinating, and is one of those textbook stories of "scientific triumph" that is so often quoted though it has so many more levels of depth to it (i.e. the simple fact that during the eclipse and before the eclipse the photographic lenses were exposed to radically different temperatures and thus had all sorts of focusing problems), not to mention all of the rather interesting problems leading up to the observation (World War I plays into it in some way, if I recall), etc. etc. Its benefit to the predictive power article is that it was indeed a prediction of Einstein's and the confirmation -- though disputed both in fact and interpretation by many scientists -- was more generally taken as a confirmation of his theory's accuracy and has certainly been retrospectively re-tooled backwards towards that end. Which is to say, I think it makes an okay little example (one which is pretty easy for a non-historian and non-scientist to understand -- Einstein predicted X, they then saw X), even if things were actually more complicated (as they always are). --Fastfission 15:52, 6 Jun 2005 (UTC)

There's no need for a disputed template. It can be modified a bit to be just fine -- it doesn't have to imply that this was the end-all truth of it. I've generally seen the case come up in science studies not so much as a case of selective interpretation but as a case of the ambiguity of interpretation, a disputation of the standard "Einstein's triumph" story, and giving a little more credit to the discontent of the anti-Einsteinians. I don't think it's the "worst example" to use -- it is just a common, overly simple example, but properly qualified I don't think it should bother anyone too much. --Fastfission 15:52, 6 Jun 2005 (UTC)

Okay, I changed it a little bit. I mentioned that this is a common example in scientific textbooks (which anybody strictly concerned with history should see as a red flag), and that it was heralded publicly (i.e., by non-scientific sources) as a confirmation, and that at the time, though, many scientists disputed both the fact and interpretation of the results. I think this should clear it up a bit. A more thorough discussion of the eclipse would warrant its own article -- as a quick little example I think it works fine here, and allows us to use that wonderful newspaper clipping on the right! I think this qualifications should please those who are worried about the details (which are not really the point of this article as it stands). Shifting it onto a "public perception" gets us out of having to be overly detailed about the historical details in this instance. You might also want to look at all which Wikipedia currently has on this topic with these things. If you'd like to propose another historical example, please feel free, however I doubt you'll find any which are indisputable (such is the case with this subject). --Fastfission 15:57, 6 Jun 2005 (UTC)

Actually, I added a whole other paragraph. We can turn this into a nice little parable. First paragraph: science textbooks give as an example of this. Second paragraph: but science studies shows us why this narrative and even this concept might be problematic. Now it is an easy-to-understand example coupled with a little bit of critique! Everybody is happy and it serves two useful purposes. --Fastfission 16:16, 6 Jun 2005 (UTC)


OK cool, nice to know other Wikipedians are in the field as well. The disputed template was a bit OTT admittedly but there doesn't seem to be one that says "this article is a little bit misleading". Anyway, the extra clarifying paragraph you've put it is a good start, though a little wordy. I might add some more relevant facts to give it some extra weight, when I get hold of my books again at the end of this week.
As an alternative example - Dmitri Mendeleev's prediction of elements that had not yet been discovered is probably a prime candidate, also the naturalist (whose name I have forgotten, might have been Darwin) who spotted extraordinarily long flowers in the jungle and postulated the existence of an insect with a very long proboscis in order to pollinate it, might well be simpler ones, but as you say, these confirmations may also have been highly disputed in their day as well... Qwghlm 16:26, Jun 6, 2005 (UTC)
Those are good examples too, I'll add them in, though I like the Einstein one especially now that it has been problematized and shows a problem with the entire concept. --Fastfission 01:40, 9 Jun 2005 (UTC)
This is why I love Wikipedia :) --Unclevortex 12:27, 9 Jun 2005 (UTC)

Err, now there's too much on the eclipse

[edit]

The eclipse was meant to just be an example of sorts. I don't think we need all of the details in this article. Perhaps someone could extract them and put them into an article on the Observations of the solar eclipse of 1919? --Fastfission 20:31, 13 July 2005 (UTC)[reply]

Factual details (such as names, dates and figures) were necessary to back up the statements made - before the edits I made it didn't stand up on its own. I don't think the example is too long - it's four paragraphs, a screen or so's worth of information, the article as a whole is well under what counts as 'large', and Wikipedia is hardly running out of diskspace, is it? The entire point of the science studies addendum was that historical accounts of science are often simplified and are only properly explained when the details and caveats are examined. You can, if you like, trim it down, but I fear by doing so you'll only get people asking "well, what exactly did happen?" several months down the line. Qwghlm 22:04, July 13, 2005 (UTC)
The statements don't have to stand up on their own in this article, though -- they can be stated and more thoroughly defended and explained somewhere else. I think they are better suited in a more specific article which we can reference (that is, link to explicitly) from this article. It's not really about physical length, it is about the reader saying, "dang, I thought this was an article about prediction, not about a specific eclipse? Why are they telling me the arc lengths? What the heck is an arc length?" and so forth; this article should use it as an example and take a number of things for granted -- one doesn't need to know the specifics of the measurements to understand the role the observations play in this particular article. Plus, the 1919 eclipse needs its own article anyway. :) --Fastfission 23:22, 13 July 2005 (UTC)[reply]
Plus, it'll give us a good place to use these neat eclipse pictures I found! ;) --Fastfission 23:24, 13 July 2005 (UTC)[reply]
Okay, you've convinced me, go ahead and do it (check first such an article doesn't already exist with a similar title - I've just realised there might be one), and leave a single paragraph here which briefly summarises the 1919 eclipse. Qwghlm 14:28, July 14, 2005 (UTC)

Sentence needs fixing

[edit]

The following sentence, from paragraph 2 of the 1919 eclipse section, needs to be fixed, but I'm not sure what it means so I can't fix it:

Two telescopes were used at Sobral, one recorded 8 plates recorded a mean deviation from the norm of 1.98″ of arc (1 ″ = 1/3600th of a degree), and the other 18 plates with a mean deviation of 0.86″

Is it that, at one telescope, 8 plates were used which recorded a mean deviation etc.? As currently stated it makes no sense. Thanks. Chick Bowen 20:20, 10 October 2005 (UTC)[reply]

Sorry, my bad. Have corrected. Is it any better now? Qwghlm 22:57, 10 October 2005 (UTC)[reply]
Yes, very clear. Chick Bowen 00:29, 12 October 2005 (UTC)[reply]

Eclipse AGAIN

[edit]

Holy smokes.

For one thing, what does this one example have to do with predictive power? No one can think of a better example to use? Like, oh I don't know, Neptune?!

For another, its generally accepted that the story that Eddington fudged the data is simply a well-repeated myth. The article here essentially states that the the myth is true, and there's no discussion about counter viewpoints. These counter-arguments appear to be extremely solid, see this entry in the Philip Ball blog for instance. Here's the cogent quote, "In any event, a modern analysis of these plates in 1979 shows that, taken together, they do support Einstein’s prediction rather well, and that the original teams made assumptions in their calculations that were justified even if they couldn’t be conclusively supported at the time."

(correct URL: Arthur-eddington-was-innocent)

All of this would be fine if this were an article on the eclipse measurements, but here it's nothing more than an overly long, poorly supported, counterexample.

I argue it should be removed entirely.

Maury 15:59, 12 September 2007 (UTC)[reply]

Off Topic?

[edit]

The name of this article is "Predictive Power," but the text mostly seems to consist a critique of the Eddington eclipse experiment, and a very superficial critique at that. The entire article seems to be off topic.

If the presence of the eclipse experiment means that the discussion should be torqued around so fully that the main focus of the article seems to be critiques of the Eddington experiment, it is badly out of place.

I rewrote the article, greatly shortening the discussion of the Eddington eclipse and moving it further back to a slightly less prominent position in the article, and I reduced the critique of the experiment to a minimum, giving a reference link to The Golem (ISBN 0521477360) for people who want more details of the critique. Geoffrey.landis 23:17, 10 November 2007 (UTC)[reply]

It is probably worth having a separate article on just the Eddington experiment. There is enough history and subsequent analysis and myths that need correcting, to justify it. I learnt about all this in the late-1990s, and I was unaware that opinions had changed again (see the 2006 and 2007 papers), or shifted slightly (not sure which). Also "observations made by the astrophysicist Arthur Eddington" oversimplifies it. It was two expeditions and many people involved. I believe Eddington went to Principe, and someone else went to Sobral (or maybe it was the other way around). And then the results were analysed when they got home. Carcharoth (talk) 20:30, 26 January 2008 (UTC)[reply]
A separate article on the experiment does sound reasonable. Geoffrey.landis (talk) 04:28, 27 January 2008 (UTC)[reply]
I'll try to start something at some point. The best account on-wiki seems to be Tests of general relativity#Deflection of light by the Sun. The article mentioned there, Kepler problem in general relativity, turns out to be a general article about the mathematics of the general problem. I'm wondering what the best title for the experiment would be. Category:Physics experiments has articles closest to what I'm thinking of (though the Category:Experiments looks like it needs re-organising). I'm thinking either "experiment" or "expedition". We have Category:Scientific expeditions. The article could also focus on the eclipse, though that distracts from the expedition and experiments. Seeing as the article should cover the results and the impact and the later history, it should be "experiment". Maybe Eddington experiment and/or 1919 Solar Eclipse Expeditions? There is also (though less widespread) the Dyson-Eddington experiment and Dyson-Eddington-Davidson experiment. Which name would be best do you think? Carcharoth (talk) 05:04, 27 January 2008 (UTC)[reply]

Note: Arthur Stanley Eddington, Frank Watson Dyson, Charles Rundle Davidson [1]. "F. W. Dyson, A. S. Eddington, and C. Davidson, Phil. Trans. Royal Soc. London. Series A 220 (1920) 291-333". Might get a new biographical article out of this as well! Carcharoth (talk) 05:12, 27 January 2008 (UTC)[reply]

Eddington experiment

[edit]

A note (following on from the above) that the main content on the eclipse and the experiment now appear to be at the following three locations:

Carcharoth (talk) 14:23, 27 May 2019 (UTC)[reply]

Definition?

[edit]

There appears to be no definition of the term "Predictive power" in the article. I left reading it with having to determine the meaning by abstracting from the given examples. — Preceding unsigned comment added by 77.0.213.82 (talk) 23:37, 1 January 2020 (UTC)[reply]