Transliteration is a type of conversion of a text from one script to another that involves swapping letters (thus trans- + liter-) in predictable ways, such as Greek ⟨α⟩ → ⟨a⟩, Cyrillic ⟨д⟩ → ⟨d⟩, Greek ⟨χ⟩ → the digraph ⟨ch⟩, Armenian ⟨ն⟩ → ⟨n⟩ or Latin ⟨æ⟩ → ⟨ae⟩.
For instance, for the Modern Greek term "Ελληνική Δημοκρατία", which is usually translated as "Hellenic Republic", the usual transliteration to Latin script is ⟨Ellēnikḗ Dēmokratía⟩, and the name for Russia in Cyrillic script, "Россия", is usually transliterated as ⟨Rossija⟩.
Transliteration is not primarily concerned with representing the sounds of the original but rather with representing the characters, ideally accurately and unambiguously. Thus, in the Greek above example, ⟨λλ⟩ is transliterated ⟨ll⟩ though it is pronounced [l], ⟨Δ⟩ is transliterated ⟨D⟩ though pronounced [ð], and ⟨η⟩ is transliterated ⟨ē⟩, though it is pronounced [i] (exactly like ⟨ι⟩) and is not long.
Conversely, transcription notes the sounds rather than the orthography of a text. So "Ελληνική Δημοκρατία" could be transcribed as [elinikí ðimokratía], which does not specify which of the [i] sounds are written with the Greek letter ⟨η⟩ and which with ⟨ι⟩.
Angle brackets may be used to set off transliteration, as opposed to slashes and square brackets for phonetic transcription. Angle brackets may also be used to set of characters in the original script. Conventions and author preferences vary.
Systematic transliteration is a mapping from one system of writing into another, typically grapheme to grapheme. Most transliteration systems are one-to-one, so a reader who knows the system can reconstruct the original spelling.
Transliteration is opposed to transcription, which maps the sounds of one language into a writing system. Still, most systems of transliteration map the letters of the source script to letters pronounced similarly in the target script, for some specific pair of source and target language. If the relations between letters and sounds are similar in both languages, a transliteration may be very close to a transcription. In practice, there are some mixed transliteration/transcription systems that transliterate a part of the original script and transcribe the rest.
For many script pairs, there is one or more standard transliteration systems. However, unsystematic transliteration is common.
In Modern Greek, the letters ⟨η⟩ ⟨ι⟩ ⟨υ⟩ and the letter combinations ⟨ει⟩ ⟨oι⟩ ⟨υι⟩ are pronounced [i] (except when pronounced as semivowels), and a modern transcription renders them all as ⟨i⟩; but a transliteration distinguishes them, for example by transliterating to ⟨ē⟩ ⟨i⟩ ⟨y⟩ and ⟨ei⟩ ⟨oi⟩ ⟨yi⟩. (As the ancient pronunciation of ⟨η⟩ was [ɛː], it is often transliterated as an ⟨e⟩ with a macron, even for modern texts.) On the other hand, ⟨ευ⟩ is sometimes pronounced [ev] and sometimes [ef], depending on the following sound. A transcription distinguishes them, but this is no requirement for a transliteration. The initial letter 'h' reflecting the historical rough breathing in words such as Ellēnikē should logically be omitted in transcription from Koine Greek on, and from transliteration from 1982 on, but it is nonetheless frequently encountered.
A simple example of difficulties in transliteration is the Arabic letter qāf. It is pronounced, in literary Arabic, approximately like English [k], except that the tongue makes contact not on the soft palate but on the uvula, but the pronunciation varies between different dialects of Arabic. The letter is sometimes transliterated into "g", sometimes into "q" and rarely even into "k" in English. Another example is the Russian letter "Х" (kha). It is pronounced as the voiceless velar fricative /x/, like the Scottish pronunciation of ⟨ch⟩ in "loch". This sound is not present in most forms of English, and is often transliterated as "kh", as in Nikita Khrushchev. Many languages have phonemic sounds, such as click consonants, which are quite unlike any phoneme in the language into which they are being transliterated.
Some languages and scripts present particular difficulties to transcribers. These are discussed on separate pages.