Johannes Goropius Becanus (1519-1572) was one of those brilliant erudites of the late Renaissance, a man who was not only a physician but a linguist as well. However, the study of language was still in its infancy, and even capable scholars like Grotius came up with ridiculous etymologies: for instance, that the Americas were colonized from Holland, because there were many Aztec towns with names ending on /-lan/, suggesting that the Aztec language shared the Dutch word “land”. You can read more about Grotius here.
Goropius had even greater ambitions: he argued that all languages were derived from Dutch. After all, the Dutch name for the Dutch language, diets, was derived from the Dutch words d’outs, “the oldest”. In his Origines Antwerpianae (1572) he offered many similar, odd etymologies.
It is easy to laugh about this, but we must be fair: this was the beginning of the field of comparative linguistics, which is now a well-respected field of scholarship. Whatever his fantasies, Goropius had the brilliant idea that systematic comparison was a good path to obtain knowledge. So it is a bit undeserved that “goropising” is the name given to ridiculous etymologies.
Nowadays, it is almost impossible to get away with it, at least in the Indo-European languages. Things are different, however, in the Semitic language group. One reason is that all words are derived from a triliteral root; for example, Arabian qrš means “to gnash, to crunch, to chew” or “to earn money”. Qirš can refer to a shark or a coin, qariš is means cheese, while mu-qriš refers to a person who has earned a lot of money.
Now, the Semitic languages have four s-like sounds and two t-like sounds that can become an S. That gives us six interchangeable sounds. Something similar can be said about the two k-like sounds and the G. Add that the letters R and L are in nearly every language interchangeable, and we reach the situation that you can derive any given word from nearly every other word.
If, contrary to expectation, things are still impossible, the goropiser can always look in a related language. The qrš-example above was from Arabic, but you can always take something from Hebrew, which may have retained words from a shared vocabulary that the Arabs forgot. Since qrš in Hebrew means “plank”, we might deduce that the Arabian word for shark does not mean “chew-fish” but “drifting piece of wood”.
To all this must be added the possibilities created by the fact that ancient and medieval clerks wrote Arabic and Hebrew without vowels and used letters that were not always easy to distinguish. For example, the Arabian B, T, and TH are distinguished by dots only, and the same applies to J, G, and K, to R and Z, to F and Q.
Given these problems, it is possible that someone comes up with an etymology that is absolutely credible, does not contradict the laws of language, and looks plausible, but can still be incorrect. You cannot establish whether the similarity between two words is evidence for a real connection or just coincidence. This means that in the Semitic languages, etymologies are an extremely unreliable – perhaps even unscientific – method.
If you want to goropise, take the Semitic languages.
