[Home]
This document discusses how phonetic notations are used in Eldamo. Eldamo uses some non-standard notations for several reasons. First, Tolkien also used non-standard notations, and it is generally clearer to adhere to Tolkien’s notations than that of modern linguistics. Second, the official notations for linguistic sound changes can be very complex and hard for non-linguists to understand. Therefore, I’ve adopted a simplified set of notations, in part to make it easy to convert sound change rules into a format usable by computure programs.
A lot of phonetic terminology appears in Tolkien’s writing and in this lexicon. The most common and important linguistic terms are discussed here.
Phoneme: A distinct sound in a language used for forming words. The common linguistic convention is to put phonetic representations inside brackets, [m] for the sound “m”, though Tolkien himself rarely did so in his writing. Phonemes often correspond to the letters of a written language, but not always. The written letter “m” in English generally represents the phoneme [m], but the letter “c” can represent either [k] or [s], depending on where it appears: “call” vs “cell”. Furthermore, different languages make different phonetic distinctions. The stereotyped Asian confusion between English [l] and [r] (“I rike you velly much” for “I like you very much”) comes from the fact that many Asian languages don’t distinguish those sounds (though in practice, many Asians can make the distinction when listening to or speaking English: the stereotype has a germ of truth, but is not the reality).
The International Phonetic Alphabetic (IPA): A standard way of representing sounds independently of language. In many (but not all) cases, the IPA use the same letters as in English to represent the sounds that appear in the English language, but it has other symbols to represent non-English sounds or sounds that are ambiguously represented in English writing. The Wikipedia IPA page has a good overview of IPA symbols and the sounds they represent. Unfortunately, Tolkien often used his own, non-standard notation in his phonetic descriptions. See below for a discussion of the symbols commonly used by Tolkien and the IPA sounds they represent.
Orthography: How the sounds of a language are represented in writing. Although Tolkien invented scripts for his languages, most of his writing used English letters to represent words, much like how Vietnamese can be written in either its native script or with English letters (or more accurately Latin letters, which is where modern English letters come from). Tolkien used various orthographic representations for his different languages, generally using the equivalent English letters for the sounds that were similar and combinations of letters for the sounds that differed. In some cases, he used different letters to make the languages more visually distinct. In The Lord of the Rings, he used “c” to represent the sound [k] in his Elvish languages, but “k” to represent the sound in Khuzdul, Adûnaic and Black Speech. Tolkien’s orthographic representations are more regular than most real-world languages. Once you know his orthographic rules for a given language, you can generally tell how a word is pronounced from how it is written, which is not the case with many English words.
Digraphs: A representation of a single sound (phoneme) with a pair of letters. Tolkien often used digraphs to represent sounds that did not appear in English. English also uses digraphs, such as “th” for the sound at the beginning of the word “thin” (IPA [θ]). Tolkien used this digraph to represent the same sound in Sindarin. Tolkien sometimes used digraphs to represent non-English sounds, such as “ch” for the sound appearing in the German name “Bach” (IPA [x]), not the sound “ch” as in “church” which is a sound that does not appear in his Elvish languages. Tolkien also used distinct digraphs for cases where English can be unclear. For example, he used “dh” in Sindarin for the sound [ð], which is pronounced like the “th” in English “this” or “that”; English ambiguously uses the digraph “th” for both [θ] and [ð].
Diacritic: A mark added to a written letter that (usually) modifies its sound. The most common diacritic used by Tolkien was the acute accent (á, ó), which indicates long vowels in his Elvish languages. To enforce visual distinctiveness, he sometimes used different diacritics in different languages for the same purpose. He used the macron (ā, ō) for long vowels in ancient languages and the circumflex (â, ô) for long vowels in Adûnaic and Khuzdul. Tolkien also used diacritics to clarify the proper use of a sound. Short vowels were usually unmarked, but Tolkien sometimes put the breve (ă, ŏ) over short vowels in ancient languages to clarify that they were not long. He frequently put a diaeresis (ë, ö) over short vowels in Elvish languages to indicate that they were pronounced separately instead of combining into a dipthong or becoming silent at the end of a word. This use of the diaeresis is sometimes seen in English words, such as naïve.
Vowels and consonants: These two terms are generally understood by most people, at least at a basic level. Phonemes can be classified into vowels (made with an open configuration of the vocal tract) and consonants (in which the breath is partly obstructed). Syllables in a word generally require at least one vowel, and consonants generally mark the boundary between syllables. Some sounds can blur the boundary between the two, like the semi-vowels “w” (IPA [w]) and “y” (IPA [j]). Think of the “a, e, i, o, u and sometimes y” classification of English vowels, though this isn’t exactly accurate from a phonetic perspective (when used as a vowel, the written English “y” usually represents a long “ee”, phonetically).
Consonants can be classified by their place of articulation: the location in the mouth where a sound is produced, which is often but not always indicated by the position of the tongue. Tolkien frequently organized the consonants of his languages into series based on the place of articulation. The shapes of Tolkien’s tengwar letters also indicate this: letters with similar shapes share the same place of articulation, though the exact classification is language-dependent.
Labial: A consonant produced at the front of the mouth with the lips. Typical English labials are [p], [b], [m], [f] and [v]. Tolkien often called groups of labials the “p-series”, which is how he labeled the parmatéma series (second column) in the tengwar table in The Lord of the Rings Appendix E. Labials can be further subdivided into bilabial (made with both lips) and labiodental (with lower lip and teeth), though Tolkien often omitted these distinctions.
Dental: A consonant produced with the tongue at or near the teeth. Typical English dentals are [t], [d] and [n], as well as the sounds [θ] and [ð] which appear at the beginning of English words “thin” and “this” respectively. Tolkien often called groups of dentals the “t-series”, which is how he labeled the tincotéma series (first column) in the tengwar table. Many of the sounds that Tolkien called dentals should strictly be classified as alveolar (with the tip of the tongue on the fleshy dental ridge behind the teeth) rather than pure dentals (with the tongue touching the teeth).
Postalveolar: A consonant produced with the tongue behind the dental ridge. Typical English postalveolars are “j” [dʒ], “ch” [tʃ] and “sh” [ʃ], as they appear in “judge”, “church” and “shush”. These sounds were rare in Tolkien’s languages. In the languages of The Lord of the Rings, postalveolar sounds only appear in Westron. In other writings, these sounds appeared in the Vanyarin dialect of Quenya, some Avarin languages and sometimes in Adûnaic.
Velar: A consonant produced with the back of the tongue. Typical English velars are [k], [g] and the back-nasal [ŋ] heard toward the end of “sing” and “link”. Tolkien often called groups of velars the “k-series”, which is how he labeled the calmatéma series in the tengwar table (the third column in Quenya usage). Tolkien’s languages frequently used velar sounds not commonly heard in English, such as the [x] sound heard at the end of the German name “Bach”.
Palatal: A consonant produced with the body of the tongue centered on the roof of the mouth. Pure palatals are almost non-existent in English, the only exception being the consonant “y” (IPA [j]). The only invented language described in Tolkien’s later writing that used a series of pure palatals was ancient Adûnaic, where he called it the “c-series” after the voiceless palatal stop, represented in the International Phonetic Alphabet by [c] (the earliest version of primitive Elvish also had a c-series). See, however, the discussion of palatalized sounds below.
Homorganic: A set of phonemes that have the same place of articulation. Homorganic consonants tend to be preferred in combinations. In English the nasal sounds [m] (labial) and [n] (dental) tend to be grouped with homorganic consonants. This is the basis for the distinct negative prefixes “im-” and “in-” in English words like “impossible” and “intangible”: [m] is combined with other labial sounds like [p], while [n] is combined with dentals like [t] (and [ŋ] with velars, see below). Homorganic nasals are similarly grouped in Tolkien’s Elvish languages.
Allophones: Different phonemes that are considered the same sound by speakers of a language. As noted above, [r] and [l] are allophones in several Asian languages (such as Japanese). The dental nasal [n] and the velar nasal [ŋ] are allophones in English, with the sound [ŋ] appearing only in combination with velars “g” and “k” such as “sing” and “link”. An English speaker would likely decompose “nk” and “ng” into “n+g” and “n+k”, but in practice they are generally pronounced with an [ŋ], especially at the end of words. Elsewhere an English “n” is usually pronounced [n]. However, the dental and velar nasals [n] and [ŋ] are distinct phonemes in the pronunciation of Classical Quenya, with [ŋ] even appearing as a distinct consonant at the beginning of some words.
Most linguistics use slashes to distinguish the sounds as they are understood by a language’s speakers, versus the real phonemes as they are actually pronounced. Compare English /n/ versus its allophones [n] and [ŋ]. Unfortunately, such fine distinctions can be difficult to deduce for Tolkien’s languages. This lexicon generally uses brackets around phonemes in all cases.
In addition to the place of articulation, consonants can be classified by the various methods used to produce the sound. These methods can combined with the place of articulation to classify a phoneme, such as the voiced velar stop [g] or voiceless labial spirant [f] (more accurately labio-dental).
Voiced and Voiceless: One of the most basic distinction in consonants, voicing indicates whether a sound is produced with or without resonating the vocal cords. Many English sounds are distinguished by whether or not they are voiced: voiced [v] versus voiceless [f], voiced [g] versus voiceless [k]. Since voicing occurs in the throat, its presence or absence can be readily combined with other methods of pronunciation.
Stop: A consonant produced by stopping and releasing the breath. Together with voicing (or its lack), such stoppage produces various consonants in the different places of articulation: labial [b] and [p], dental [d] and [t], velar [g] and [k]. Comparing these sounds is a good way to understand both the effects of voicing and the different places of articulation.
Spirant (Fricative): A consonant produced with “vocal friction”, with partial closure producing turbulent air flow with a somewhat hissing sound. Modern linguists typically call such sounds fricatives, but Tolkien generally used the older term spirant. Friction can also be combined with voicing or the lack thereof: labial [v] and [f], dental [ð] and [θ] (heard in “this” and “thin”), alveolar [z] and [s].
Nasal: A consonant produced by directing the flow of air through the nose instead of the mouth. The position of the tongue in the mouth still influences such sounds: labial [m] versus dental [n] versus velar [ŋ] (heard toward the end of “sing”). Nasals are usually voiced, but can sometimes be unvoiced.
Other Sounds: Although voicing, method and place of articulation can be used to classify many sounds, others defy simple descriptions, such as liquids [r], [l] or semi-vowels [w], [j] (English “y”). Such phonemes often have ambiguous places of articulation, which lets them combine more easily with a variety of other sounds.
In some cases, the basic method of articulation of a sound can be further modified to produce a different consonant. These modifications are phonetically represented by superscript letter or a diacritic mark.
Labialized: A consonant modification involving rounded lips. Labialization is often the result of a consonant followed by a “w” [w] sound. A notable English (and Quenya) example is “qu”, phonetically [kw] or [kʷ]. Several Quenya words also begin with labialized [nw] or [nʷ], such as Quenya nwalmë (the name of tengwa #20). In Classical Quenya this sound was labialized [ŋʷ]. These palatalized sounds are written in Quenya with the quessetéma series, the fourth tengwar column in Quenya usage. Labialization is represented in the IPA with a superscript-w: [ʷ].
Palatalized: A consonant blended with a following “y” sound (IPA [j]). In English, the “ty” and “ny” sounds heard at the beginning of “tune” and “new” are examples of palatalized consonants. Palatalized consonants are written in Quenya with the tyelpatéma series, which is the first tengwar column with an added diacritic mark (two underposed dots) to indicate the following y-sound. For example, 1Î = “ty” [tʲ]. In the IPA, palatalization is represented with a superscript-j, [ʲ] as in [tʲ] or [nʲ], because the IPA uses the symbol [j] for the sound of the English consonant “y”.
Aspirated: A consonant followed by an h-like puff of air. Aspiration in the IPA is represented with a superscript-h [ʰ]. In English, the consonants “p”, “t” and “k” are generally aspirated at the beginning of words before a vowel, but unaspirated at the end of words: compare “tin” versus “at”. If you hold your hand in front of your mouth, you can feel the difference in the flow of air between the two t’s. In English, [tʰ] and [t] are allophones of /t/, but they are distinct phonemes in some of Tolkien’s languages, such as Primitive Elvish and Khuzdul.
The symbols used by Tolkien to represent various sounds can be rather complex. Not only did he use different symbols for different languages, he used different symbols in different time periods, for example using “ñ” for the palatal nasal in his early writings (PE12/15), but for the velar nasal in his later writings (PE18/82). This lexicon normalizes spelling of its word-entries to the most common orthographic representations Tolkien used for a given language and time period. Thus, the lexicon uses “qu” to represent [kw] or [kʷ] when discussing Quenya words from the time period of The Lord of the Rings, but uses “q” for the time period of the Etymologies and the early Qenya Lexicon, while using “kw” when representing Primitive Elvish. These different representations of the same sound are simply the conventions followed by Tolkien at various points in his writing.
However, when specifically discussing phonetics, this lexicon standardizes on IPA symbols to remove ambiguity, putting IPA notation in brackets next to the symbols used by Tolkien in the cases where the intended sound might be unclear. This requires guesswork in some cases, since it isn’t always certain what sound Tolkien meant by a particular symbol. The list below includes the most common specialized phonetic symbols used by Tolkien in the Etymologies of the 1930s and later writings, together with their IPA representations. For earlier works, see the list of symbols compiled by Arden Smith on PE16/10-13.
Note that symbol [ʒ] has a different meaning in IPA than Tolkien’s use of it. In the IPA, [ʒ] is used for the voiced post-alveolar spirant, the voiced equivalent of [ʃ] (English “sh”). This IPA sound appears in English as the “z” of “azure” or the “s” of “fusion”. However, Tolkien consistently used the symbol ʒ for what he called the “voiced back spirant”, represented in IPA by [ɣ] (Tolkien’s abnormal use was actually based on a different but similar looking Old English letter ȝ “yogh”). This sound is not found in English. You can produce it by starting with the voiceless velar spirant “ch” of “Bach” (IPA [x]) and adding voicing to the sound. Fortunately for English speakers, this sound was only used in prehistoric Elvish languages, and disappeared from Quenya and Sindarin.
This lexicon uses some non-standard notation to describe sound changes. Its XML data model is designed for computer processing, so it describes sound changes in a form more amenable to computer languages. These patterns are inspired by the notation for Regular Expressions, but are not identical to it.
To describe the phonetic change, this lexicon uses a set of matching rules to indicate which sounds are modified by a given rule. Each matched phoneme is represented by a single character, with the possible addition of diacritic marks such things as vowel lengthening [ā], as well as markers for aspiration [ʰ], palatalization [ʲ] and labialization [ʷ].
Each phoneme+diacritic requires an exact match. Groups of phonemes in braces ({}) must match exactly one of the elements in the braces. For example, [m{ppʰ}] would match an [m] followed by either a [p] or [pʰ].
A capital V matches any vowel. Similarly, a capital C matches any consonant. Vowels may further marked as short V̆ or long V̄; if unmarked, the pattern matches any vowel. Groups of consonants may also be marked with diacritics: [{ptk}ʰ] would match any of the specified aspirated voiceless stops: [pʰ], [tʰ] or [kʰ]. A capital X matches any phoneme: vowel or consonant. Finally, a capital S matches a syllable: a vowel or diphthong, possibly preceded and/or followed by one or more consonants.
A dash in a pattern indicates the position of the pattern inside a word. If a dash appears at the end of the pattern, it only matches initial sounds: [m-] would only match an m at the beginning of a word. Similarly, if a dash appears at the beginning, it only matches final sounds: [-m] would only match an m at the end of a word. Finally, if dashes appear both before and behind, it only matches medial sounds: [-m-] would only match an m in the middle of a word.
In phonetic derivation, one phonetic pattern changes to another, indicating how the phonemes change. Each phoneme in the first pattern changes to the phoneme in the second pattern that has the corresponding position. For example, the derivational rule [m{ttʰ}] > [n{ttʰ}] indicates that the matching values [mt] and [mtʰ] would change into [nt] and [ntʰ], with the preceding [m] changing to [n] but the following [t] or [tʰ] remaining the same. Similarly, the derivational rule [m{ttʰ}] > [m{ppʰ}] would indicate the [mt] and [mtʰ] change to [mp] and [mpʰ], with the preceding [m] remaining the same and the following [t] or [tʰ] changing to [p] or [pʰ], respectively.
In cases where a phoneme disappears, the target pattern indicates the vanished phoneme with a [ø]. For example, the rule [VɣV] > [VøV] indicates that [ɣ] disappears between vowels. In cases where a phoneme is added, the position where the new phoneme will appear is also marked by a [ø]. For example, the rule [øṇd-] > [and-] indicates that the vowel [a] is added before the syllabic [ṇ], which then becomes non-syllabic. In cases where the transformation is unambiguous, the ø-sign can be omitted: [ṇd-] > [and-].
Transformations that alter one of the surrounding characters are indicated by diacritic marks. For example, the rule [VɣC] > [V̄øC] indicates [ɣ] disappears between a vowel and consonant, but the vowel is lengthened. Having V̄ > V̆ in a pattern would indicate a long vowel is shortened. Such diacritic markers can apply to consonants as well: the rule [{ptk}ʰ{ptk}] > [{ptk}{ptk}ʰ] indicates that when an aspirated voiceless stop is followed by an ordinary voiceless stop, the aspiration moves to the end.
In cases where phonemes change position (metathesis), phoneme groups in braces are marked with a numeric subscript to indicate which group moves where. For example, [{lr}₁{tk}₂] > [C₂C₁] indicates that where an [l] or [r] is followed by a [t] or [k], they switch position. Subscripts can also be used for other transformational dependencies. The rule [V₁Cøl] > [V₁CV₁l] indicates that a vowel is added between the consonant and a following [l], and that the value of the new vowel is identical to that of the preceding vowel.
Finally, a series of patterns may be combined into a single rule by separating each pattern by a vertical bar “|”. Each component in the first pattern changes into the corresponding component in the second pattern. Consider the three rules [ṃb-] > [umb-], [ṇd-] > [and-], [ŋ̣g-] > [iŋg-]; these can be combined into the single rule [ṃb-|ṇd-|ŋ̣g-] > [umb-|and-|iñg-]. The “|” can also separate groups in braces: the rule [-{mb|nd}-] > [-{mm|nn}-] would indicate that medial mb and nd change to mm and nn.
TBD: *-patterns, stress markers