It originally meant "bridge" in the Etymologies, but was apparently switched with iant
Primitive elvish
yalta
root. yoke
Quenya
yalta
noun. yoke
Noldorin
iant
noun. yoke
ianw
noun. (?) yoke
Sindarin
ialt
noun. yoke
ianu
noun. (?) yoke
It originally meant "bridge" in the Etymologies, but was apparently switched with iant
iant
yoke
iant (bridge), pl. iaint, coll. pl. iannath
iant
yoke
(bridge), pl. iaint, coll. pl. iannath
Beware, older languages below! The languages below were invented during Tolkien's earlier period and should be used with caution. Remember to never, ever mix words from different languages!
Early Quenya
yalta
noun. yoke
yalte
noun. yoke
yarta
noun. yoke
Early Primitive Elvish
yḹta
noun. yoke
Qenya
yanta
noun. yoke
The Etymologies of the 1930s had ᴹQ. yanta “yoke” derived from primitive ᴹ✶yantā under the root ᴹ√YAT “join” (Ety/YAT). However, in Tolkien’s later writings, Q. yanta was translated as “bridge” instead (LotR/1123).
Conceptual Development: In the Qenya Lexicon of the 1910s Tolkien had ᴱQ. {yatra >>} yarta “yoke” under the early root ᴱ√ẎATA “join” (QL/105), as well as ᴱQ. yalta “yoke” derived from ᴱ✶yḹta under the early root ᴱ√YḶTḶ (QL/106). In the Gnomish Lexicon also from the 1910s, Tolkien had ᴱQ. yalte as cognate to G. galt “yoke” (GL/37).
Neo-Quenya: Of the options above, I think ᴺQ. yalta is the best choice for a Neo-Quenya word for “yoke”, as a cognate to ᴺS. ialt. For example, Helge Fauskanger used both yanta and yalta for “yoke” in his NQNT (NQNT).
Middle Primitive Elvish
yantā
noun. yoke
Gnomish
galt
noun. yoke
The Etymologies of the 1930s had N. iant “yoke” derived from primitive ᴹ✶yantā under the root ᴹ√YAT “join” (Ety/YAT). Deleted forms iad and [?]iannas were written immediately after it (EtyAC/YAT). In Tolkien’s later writings S. iant was translated as “bridge” (SA/iant; SD/129; WJ/333), as opposed to N. ianw “bridge” from The Etymologies (Ety/YAT).
Conceptual Development: The Gnomish Lexicon from the 1910s had G. galt “yoke” with cognate ᴱQ. yalte (GL/37), which was probably derived from the early root ᴱ√YḶTḶ from the contemporaneous Qenya Lexicon which also had ᴱQ. yalta “yoke” as a derivative (QL/106). In Gnomish of the 1910s ancient initial y became g, which explains G. galt vs. ᴱQ. yalta.
Neo-Quenya: For purposes of Neo-Sindarin, I would revise G. galt to ᴺS. ialt “yoke” as a cognate to ᴺQ. yalta, since S. iant is used for “bridge”.