A reference to a primary source would be, for example, a reference to Newton's "Philosophiæ Naturalis Principia Mathematica" or to Darwin's "On the Origin of Species". An article on the history of evolutionary thought would be remiss if it does not include Darwin's book and instead only referred to books discussing Darwin's work. Like the children's game "telephone", only relying on secondary (and then tertiary, then quaternary) sources can amplify noise.
You then asked about "examples of original research published in the Encyclopedia of Physics". There are two different types of research involved here. One is "original to the field of physics". Physics has a well-established mechanism for publishing and disseminating work, and that is not an encyclopedia.
The other is the research needed to reconcile and synthesize multiple viewpoints into a well-constructed whole. A non-fiction piece for the New Yorker likely entails new research (even an interview is new research), and that's the type of research which goes well with an encyclopedia. Take a look at Wikipedia's entry for "History of the Encyclopædia Britannica" with comments like "40,000-word hagiographic biography of George Washington" and "Dr. Thomas Thomson, who introduced the first usage of chemical symbols in the 1801 supplement". Thomas Young translated the Rosetta Stone and in his WP page is written "[s]ome of Young's conclusions appeared in the famous article "Egypt" he wrote for the 1818 edition of the Encyclopædia Britannica."
Or take Asimov's Biographical Encyclopedia of Science and Technology. The entire encyclopedia was written by Asimov, who commented "I alone have done every bit of the necessary research and writing; and without any assistance whatever, not even that of a typist." Right there in the text it says "research."
Do you think these examples of the research which goes into encyclopedia articles aren't actual research? If not, why not? Or are these simply not encyclopedias?
As to "if Britannica is paying"... do you think Harry Houdini, Albert Einstein, Marie Curie, Sigmund Freud, Henry Ford, Leon Trotsky, Arthur Eddington, Lord Kelvin, Humphry Davy, John Stuart Mill, Thomas Malthus, and yes, even Isaac Newton, were contributors to EB mostly because they were being paid for their work? Most certainly not! (How much would Ford's time cost?) From what I've read, other factors were because they wanted to contribute to a collection of knowledge, and because of the prestige.
