I was also heading toward the idea that the original aluminum wiring had been "pigtailed" with copper connections.
I had to do this with every outlet and fixture in an old mobile home I owned for a while.
The process involved attaching a copper stud to the existing aluminum wire with a spring-loaded wire nut, well greased to keep oxygen off the aluminum, then make the copper connection to the fixture in the normal way.
This was a royal pain in the ass, but was designed to prevent fires starting from the aluminum wiring, see NOTE.
I could see the grease applied getting all funky over time and not protecting the aluminum anymore, thus destroying the "goodness" of the Al-to-Cu connection.
Terry, 230RN
NOTE:
It took a while to realize this, but house fires had been due to alumimum wiring.
1. Al has a much higher coefficient of expansion than Cu.
2. Al forms an oxide layer almost instantly, and aluminum oxide is a great insulator. In fact, aluminum oxide ceramic insulators are used in high-voltage applications.
In the normal Cu wiring, turning down the screws to make the connections is relatively permanent and stable, and since the connection is so good, I2R heating is minimal since R is so low.
But with aluminum under that same screw, any heating results in expansion of the aluminum, and with cooling after the appliance is shut off, the aluminum contracts, reducing the dimensions of the aluminum and therefore loosening the "goodness" of the connection and allowing oxygen to contact the aluminum and form an insulating layer. Not much, but it's there.
This can result in a runaway condition, where every time the appliance is turned on and off, the situation gets worse. Not catastrophically and suddenly, but gradually over long periods. Ultimately, in some cases causing sparking or overheating and possibly a fire.
It took a long time and a lot of fires for this to finally come to light
The "fix' was to use a spring-loaded wire nut between the Al wiring and a short copper stub going under the appliance screws. The spring loading compensated for the expansion/contraction, and heavy greasing kept the oxygen off the aluminum portion of the connection.
That's the way I understand it, but I've heard of various alternate explanations for the aluminum-wired house fires.