According to Russell, are sense data NOT the only things immediately known, or are they not purely mental?

Question

In Russell's Problems of Philosophy, in refutal Berkeley's argument for idealism -in Russel's words "whatever can be immediately known must be in a mind"- he states:

Berkeley was right in treating the sense-data which constitute our perception of the tree as more or less subjective, in the sense that they depend upon us as much as upon the tree, and would not exist if the tree were not being perceived. But this is an entirely different point from the one by which Berkeley seeks to prove that whatever can be immediately known must be in a mind.

From this I can only conclude that either:

1. whatever can be immediately known includes something other than sense-data.

2. sense-data is something not purely mental.

Now if 1 is correct, then the question remains as to what is there besides sense-data that can be immediately known. (Besides "internal" things, i.e: memories, dreams, etc.. We can't take these into account since the point of idealism was to prove that even things existing "outside" are ideas).

But if 2 is correct, then what is the real meaning of sense-data? To explain my current conception of sense-data: I take the "red" sense-data to be the "redness" that I see. and if I were to hear "red" (as could happen to people with Synesthesia) it would no longer be the same thing ("red"), even if it is caused by the same physical thing and there would be a one-to-one correspondence between the "red_"s that I hear and the "red"s that I see.
Hence I take "red" to be something purely personal and mental. So if it is not purely mental (since we're supposing 2) it must have some other meaning as Russell uses it. What is that meaning?

Answer 1

The SEP chooses to put it like this

it was intended only to denote that which we are directly aware of in perception. The term's meaning was supposed to be neutral between direct and indirect realist theories of perception, so that it was not to be assumed either that sense data must by definition be mind-dependent or that they must be mind-independent

So 1 is false, but 2 is not necessarily true. Something can be dependent in common on several things at once, and sense-data depend both on the sense and on the origin of the sensation. There has to be some point in the process of apprehension where data enters the mind. The 'stuff' at that point is sense-data. It is perceived directly, but it originates partly outside the mind, or it could not be entering.

To assume entirely otherwise -- that nothing enters the mind -- puts the tree in the mind, and takes you down Berkeley's path to where there is only one mind. If I and you both perceive the tree, and the tree is in the mind, one or both of us does not exist.

This conception is from before modern theories of probing perception suggested that perceptual information flows both ways, to meet in the middle. The mind makes the guess that the red is likely, and queries the sense for validation or rebuttal. In that case the 'red' is entirely inside the mind, but the test criteria for redness still uses facts from outside of it, and those become the sense-data, but not exactly in Russel's sense.

Theories that do not allow for the mind taking an active role in apprehension, by asserting assumptions about qualia, generally fail to explain why we really perceive the exact same physical data in different ways in different contexts.