You have two or three people mixed up here. It's compounded by the several Alexander-Mathis marriages. Thomas J. Alexander, the son of William Alexander, was not the same as Martha Mahalia's father, Thomas C. Alexander, born about 1803, Logan County, KY. Martha's mother was Elizabeth McMillin, Thomas C's cousin. Thomas J. was born about 1812, probably Logan County, and his parents were Thomas C.'s much older brother, William, and probably Hettie (real name probably Esther) Hinton.
Also, Matthew's father was James Sr, but James Sr's father was not Moses. DNA evidence proves this. If there is any kinship between James Sr and James the weaver, the lines probably don't join for several thousand years back.
If the Elizabeth Jane Mathis in your file is the daughter of Leonard Mathis, I am related to this family through the Alexanders, McMillins, and Mathises.