This seems to be a question whose answer depends on whom you ask. Maybe we can come up with a final answer?
It is known that Poincaré, at least, invented something that can be called Poincaré-Birkhoff-Witt theorem (PBW theorem) in 1900. Okay, it was a version of the PBW theorem that required a basis of the Lie algebra, and it was formulated over $\mathbb R$ or $\mathbb C$ only, but this is not unexpected of a 1900 discovery, and generalizing the statement is pretty straightforward from a modern perspective (generalizing the proofs, not so much).
What is unclear is whether Poincaré gave a correct proof. Poincaré's proof appears in two places: his paper
Henri Poincaré, Sur les groupes continus, Transactions of the Cambridge Philosophical Society, 18 (1900), pp. 220–255 = Œuvres de Henri Poincaré, vol., III, Paris: Gauthier-Villars (1934), pp. 173–212,
and the modern paper
I have no access to the former source, so all my knowledge of the proof comes from the latter.
Ton-That and Tran claim that Poincaré's proof was long misunderstood as wrong, while in truth it is a correct, if somewhat incomplete proof. The incompleteness manifests itself in the fact that a property of what Poincaré called "symmetric polynomials" (and what we nowadays call "symmetric tensors") was used but not proven. However, in my opinion this is not a flaw: This property (which appears as Theorem 3.3 in the paper by Ton-That and Tran and is proven there in an overly complicated, yet nice way) is simply the fact that the $k$-th symmetric power of a vector space $V$ over a field of characteristic $0$ is generated by $k$-th (symmetric) powers of elements of $V$. This fact is known nowadays and was known in 1900 (I think it lies at the heart of umbral calculus).
What I am not sure about is the actual proof of PBW. Since I don't have the original Poincaré source (nor, probably, the understanding of French required to read it), I am again drawing conclusions from the Ton-That and Tran paper. My troubles lie within this paragraph on pages 277-278:
"The first four chains are of the form
$U_1 = XH_1,\ U'_1 = H'_1Z,\ U_2 = YH_2,\ U'_2 = H'_2T$,
where each chain $H_1$, $H'_1$, $H_2$, $H'_2$ is a closed chain of degree $p – 1$; therefore by induction, each is the head of an identically zero regular sum. It follows that $U_1$, $U'_1$, $U_2$, $U'_2$ are identically zero, and therefore each of them can be considered as the head of an identically zero regular sum of degree $p$."
Question: Why does "It follow[] that $U_1$, $U'_1$, $U_2$, $U'_2$ are identically zero"? This seems to be equivalent to $H_1 = H'_1 = H_2 = H'_2 = 0$, which I don't believe (the head of an identically zero regular sum isn't necessarily zero). The authors, though, are only using the weaker assertion that each of $U_1, U'_1, U_2, U'_2$ is the head of an
identically zero regular sum of degree $p$ – but this isn't obvious to me either.
What am I missing? Is this a mistake in Poincaré 1900? Or have the authors of 1 misrepresented Poincaré's argument? Has anybody else tried to decipher Poincaré's proof?
PS. I have asked more or less asked this question some months ago, but it was hidden in another question and did not receive any answer. Mailing the authors did not help either. So my last hope is a public discussion.
Best Answer
I don't know if Poincaré proved PBW in 1900, but Alfredo Capelli did it for $\mathfrak{g}\mathfrak{l}_n$ ten years before. Here is the link to Capelli's paper (in French).