It's my belief that before, let's say, the mid-1990s, "diversification" did not normally mean Markowitz diversification, but was a vague, qualitative term that meant not much more than "about a dozen stocks, not just one or two," and, depending on the situation, a choice between everything in one select sector (e.g. utilities) versus sector diversification--with sector diversification being a preference rather than an imperative. Furthermore, the word "diversification" very often was a reference to the business operations of an individual company, rather than to a stock portfolio.
I've been skimming The Intelligent Investor [1973 edition] (no, I've never read it through) trying to synthesize what Benjamin Graham meant by "diversification," and also as a proxy of how the word was generally understood in the 1970s and earlier.
In chapter 5 of The Intelligent Investor, he actually says
An example of a specific recommendation for portfolio diversification is:There should be adequate though not excessive diversification. This might mean a minimum of ten different issues and a maximum of about thirty.
I may have missed it but I didn't see any explicit guidance on how to know whether a list of stocks is "diversified."The defensive investor who follows our suggestions will purchase only high-grade bonds plus a diversified list of leading common stocks.
An example of the word "diversification" to mean "within a business" rather than 'within a portfolio:"
Although I can't find a completely explicit statement, I see a number of statements that suggest that for a "defensive investor" he supports the idea of a heavy concentration in public utilities; e.g....the acquisition of smaller firms by larger ones, usually as part of a diversification program.... the gospel of diversification of products has been adopted by more and more managements.
Finally, it seems clear to me that to Graham, diversification of any kind is purely a defensive measure, and a way of making sure your portfolio does not depart too far from the average. I see no hint at all that Graham ever saw diversification in the Markowitz sense, as a kind of active process that could boost the risk-adjusted-return of portfolios through low correlation--not even between stocks and bonds. The name "Markowitz" does not appear in the book. The word "correlation" appears only once in the portion written by Graham, and it's irrelevant:Actually, there are many other companies of quality equal to excelling the average of the Dow Jones list; these would include a host of public utilities.
I did a quick skim in an effort to see whether the concept of low correlation is there implicitly under a different name, and didn't find anything obvious.By an unfortunate correlation, during the same time stock-buying public has been developing an ingrained preference for the major companies and a similar prejudice against the minor ones....