Talk:Path dependence

From Wikipedia, the free encyclopedia

The section about driving on different sides of the road should be removed as it is invalid. The initial choice was not accidential, the two sides are not technically equivalent, not all countries that were in the British Empire drive on the left and not all countries that drove on the left were in the British Empire, and it has been proven that countries can change sides in modern times if the will is there.

Apart from the VCRs example, I think something could be said about path-dependence and motors. Many people think path-dependence is behind low development of electric motors and high development of gasoline motors and not-so-high development of diesel motors. Even more, path-dependence and QWERTY effects are related.

Contents 1 Split the article? 2 Where's W. Brian Arthur and his works? 3 Path independence in Social Choice and Individual Values 4 Technology 5 broken link

[edit] Split the article?

This should be made a disambig. Path-dependence in economics has precious little to do with path-dependence in physics. --Smack (talk) 22:52, 4 January 2006 (UTC)

I would like to revive the idea of splitting this article. While I have no formal training in economics or the social sciences, I do understand the use of the phrase path dependence in physics (particularly in thermodynamics), and I completely agree with Smack's assertion that the two concepts have little, if any, overlap.

In physics, path dependence is defined in the context of the possible states of a system. Having chosen a beginning state and an ending state, one can then say that a specific aspect of a process may rely on the path taken between those same two states. That is, the process can have different characteristics even when all the available paths begin at the same state and end at the same state. For example, the work, or energy input/output, required to compress or expand some quantity of gas from one combination of pressure and volume to a different pressure and volume depends on whether you (1) change the pressure first and then the volume, (2) change the volume first and then the pressure, or (3) change both the pressure and the volume together gradually.

On the other hand, all examples of "path dependence" given in this article describe systems with paths that end at different states. For example, the introduction mentions the competition between VHS and Betamax: due to the bandwagon effect and vendor lock-in, one path ends with VHS dominating and the other ends with Betamax dominating. Furthermore, the economics section explicitly states that the systems under discussion have multiple possible outcomes and that the actual ending state depends sensitively on events at the beginning of the path.

All of the article's examples, including the brief mention of symmetry breaking in the article's physics section, are simply examples of chaotic systems with multiple attractors. While this concept may go by many names in different fields, it is not related to the idea of path dependence as defined in physics. (At least one of the references[1] makes the comparison to chaos theory. However, it incorrectly mentions physics and also misinterprets chaotic systems as always lacking stable states.) This justifies the use of a disambiguation page with two articles: Path dependence (physics) and Path dependence (economics). JCarlos 17:42, 2 June 2007 (UTC)

It seems a good idea to split the article, although putting path dependence in the economics section really bothers me. The concept is widely usen in political science (Pierson 2002, North 1990, Kay 2005 and many more!) So basically I suggest to split the article to Path dependence (physics or math like JCarlos suggests) and Path dependence (politics) This comment was added by User:74.56.85.87 in between JCarlos' comments above. I have moved it here. Classical geographer 13:54, 9 November 2007 (UTC)

I would also be in favour of this, though perhaps would prefer Path dependence (math) to (physics), since that is more general. Inexact differentials are not used only in thermo. A similar concept occurs when analyzing control (and other) system with hysteresis; roughly, there's two kinds of memory, dynamics (time-dependent, analogy with C1-continuity for a curve), and hysteresis (path-dependent, analogy with G1-continuity for a curve). I would try to fill this in, but with the article's present condition, there doesn't seem to be much point. 219.127.154.18 12:43, 22 September 2007 (UTC)

The linked article (hysteresis) describes what is essentially a non-Markov process. It is more closely related to "path dependence" as described in most of this article — i.e. chaotic systems with multiple attractors — than to the concept of path dependence in thermodynamics. Unfortunately, both of these concepts can be described mathematically, so a "(math)" title would probably not suffice for either page after the split. On the topic of titles, however, my original suggestion for an "(economics)" title is probably not general enough for that page. Perhaps it could be titled "Path dependence (chaotic systems)".

JCarlos 05:23, 1 October 2007 (UTC)

The splitting is a good idea. But the point about "economics" being narrow might be good too. Maybe a solution would be "social science". Although "economics" might be fine since political science borrowed from there. In any case, "systems" is extremely misleading as a way to refer to, say, Qwerty-nomics etc. Typewritten 16:06, 8 November 2007 (UTC)

Path dependence (math) could equally be claimed by Path dependence (physics) and Path dependence (economics). Subjects like this benefit from articles named very specifically, so Path dependence (economics) is preferable to Path dependence (social science). Also Path dependence (history) and Path dependence (politics) cover much younger fields (I believe) so articles on them are likely to be less stable than those on economics/physics. Another usage: "Path Dependence" is used differently in finance than in economics; "Path Dependendent Options" (especially without state changes, like asian options) are just non-Markovian in the physicist's sense, but might be included within Path dependence (economics) because economics & finance both deal with money. Let's split the article immediately into economics & physics children, with further children as necessary, and let's clearly give the two separate implications (nicely explained above): In economics & history, saying that something is "path dependent" essentially means that its inexplicable by general causes & macro analysis, whereas in physics & finance it means non-markovian. --Wragge 12:36, 9 November 2007 (UTC)

Yep, I agree on "(physics)" vs "(economics)" as the most reasonable broad distinction. -- Typewritten (talk) 18:13, 10 January 2008 (UTC)

Me too. Apart from the problem of Finance, which can be explained within the (economics) page. It seems lots of people are in favour of splitting, so I guess whoever wants to take it up could start doing so. Classical geographer (talk) 09:15, 11 January 2008 (UTC)

[edit] Where's W. Brian Arthur and his works?

It is a shame that the name is not even mentioned. - Onlooker - 17 Apr 2006

[edit] Path independence in Social Choice and Individual Values

The last page (120) of Arrow's book (available online thru the "Table of Contents" External link) gives an argument for path independence, which transitivity as to the social ordering ensures. Calvin B. Hoover called path dependence the coral-reef effect. But of course foresight, discussion, argument, persuasion, and social (r)evolution remove any historical necessity (as distinct from possible tendency) of path dependence. Thomasmeeks 02:14, 5 July 2006 (UTC)

[edit] Technology

"It is not clear, however, that there is any inefficiency involved in the costs of remaining compatible with past decisions." I think that paragraf should be removed. It's pure speculation and really doesn't bring anything new to the subject at hand. It is at best simply wrong. —Preceding unsigned comment added by 130.225.194.192 (talk • contribs) 15:15, 31 October 2006

Eh. It's just a statement that the author isn't aware of any clear research that shows backwards compatibility leads to net inefficiency. This may reverse the burden of proof, but I don't think that makes it untrue. If it IS true, or at least it is true that we have yet to suggest net inefficiency, then it is notable and worthwhile.

However, I could be pressed to believe that some study private or public exists on social costs of compatibility with older standards. 24.160.241.34 (talk) 02:07, 23 February 2008 (UTC)

[edit] broken link

A link seemed not to go to a forum thread that exists anymore. So I removed this from the external links section:

Debate

• Forum Debate a lively and informative ongoing debate over open vs. closed systems, standards, formats, vendor lock-in, and path dependence as a deliberate strategy by some businesses

—Isaac Dupree^(talk) 18:44, 14 January 2007 (UTC)