Almost every popular account of dragons opens with a claim of the following shape: dragons appear in every culture, which shows something deep about the human mind, or about a memory of dinosaurs, or about snakes. The claim survives because the English word has been used to translate a great many unrelated creatures.
Chinese long, Japanese ryu, Indian naga, Mesoamerican feathered serpents, Mesopotamian mushussu, Greek drakon and Germanic hoard-worms have been rendered as dragon by translators reaching for the nearest available term. Once translated, they can be compared, and the comparison then appears to reveal a universal. It reveals the translation.
The case worth setting out in full is the Chinese one. The long is wingless, benevolent, associated with water and rain rather than fire, petitioned in drought, adopted as an imperial emblem, and never killed by a hero. The European dragon is winged, malevolent, associated with fire and hoarded gold, and exists in order to be killed. There is no shared origin, no shared form, no shared moral valence and no shared narrative function. They share an English word and a general reptilian outline.
That is not to say nothing can be said comparatively. Serpentine adversaries subdued by a god or hero are genuinely widespread in Indo-European and Near Eastern material, and the pattern connecting Indra and Vritra, Thor and the Midgard serpent, and Baal and Litan is a real object of study with real linguistic evidence behind it. But that is a specific and bounded argument about related traditions, not a claim about all dragons everywhere, and it deliberately excludes most of the creatures the popular version sweeps in.