Fourteen of the twenty-eight Arabic letters, exactly half the alphabet, appear disjoined at the opening of a number of suras. The classical commentary tradition records a wide range of proposals for what they do and returns no agreed answer, and one position long held within that tradition is that the meaning was not disclosed. Modern scholarship has added further proposals without producing a consensus either. This is not a case of two camps disagreeing; it is a case in which no reading has been established.
What would resolve this Nothing in the manuscript record settles the meaning, and it may not be settleable. What would narrow it is evidence that bears on the openings as written objects rather than as a puzzle: systematic collation of the earliest codices for whether and how the letters are set off, and a distributional study relating the sets of letters to the suras that carry them.
OpenQuranic studies, Arabic philology and early Islamic manuscript palaeographySeyyed Hossein Nasr (editor-in-chief), The Study Quran: A New Translation and Commentary