The first line of the Tao Te Ching is six characters long and no two translations of it agree. The difficulty is not that the Chinese is obscure but that it is underdetermined in ways English cannot leave open.
The character rendered as speak or tell is the same character as the one rendered as way, used in a second sense. So the line turns on a pun that is also a claim: the way that can be way-ed is not the constant way. English has no single word that is both a road and an act of speaking, so a translator must pick, and picking closes something the original leaves open. Some render it as speaking, some as telling, a few as treading or following, which yields a quite different sentence about ways that can be walked.
The word rendered constant carries its own history. In the received text it is the character for constant or enduring; in the Mawangdui manuscripts it is a different character with much the same sense, changed later to avoid the personal name of a Han emperor. That kind of taboo substitution is ordinary in Chinese textual transmission and is one of the routine ways the manuscripts and the received text can be told apart.
None of this is a reason to distrust translations. It is a reason to read more than one, and to treat the differences between them as information about the text rather than as evidence that somebody got it wrong.