skip to content

Restoring Lost Songs: Boethius' Consolation of Philosophy


This section is intended to provide a basic introduction to the processes of melodic reconstruction, although many of the points raised also apply to preparing a performance from an existing reconstruction.

1) Read the Poem

Reading the poem is the first step in any song reconstruction since the poetry affords a rich set of possibilities for melodic rendering. The more textual features that can be identified in any given poem, the greater the number of options that can be explored in reconstruction.  To begin with the poetic text is also to reconstruct stages in the medieval creative process, seeking to establish how a poem was understood before passing to the conventions that might have been applied in melodic realisation. The task of understanding the poem may itself be separated into distinct stages.

            i) Establishing a Text

Boethius’ De consolatione philosophiae was not easy to understand for medieval scribes as witnessed by errors in the copying process.  Transcribing the poetic text directly from the source risks re-inscribing mistakes that would have been recognized as such by more educated readers at the time.  It is simplest for practical purposes to begin instead from modern editions: Moreschini’s edition is now standard, although Bieler’s edition remains highly respected.  See here for details.

            ii) Comprehension

Finding a good modern translation is an essential first step for all but the best Latinists due to the complexity of the text.  Recommended translations can be found here.  Understanding the sense of an individual metrum depends in part on appreciating its place within the overall structure of the De consolatione philosophiae.  It is therefore necessary to understand how the metrum relates to the preceding prose passage and where the metrum occurs within the narrative of the five books.

            iii) Genre

Boethian metra were not written within established song genres, if only for the simple reason that none were composed in strophic forms.  Even so, many of the metra draw on characteristics of laments, hymns, love songs, nature songs, astronomical songs, moralistic songs, ballads etc.  Awareness of informing song styles can help by directing attention to certain models for reconstruction and suggesting an overall approach within performance. At the same time it has to be remembered that the metra are not uniform.  Several poems weave together multiple stories (see Bella bis quinis, IV:7) or display a shift in attitude within the song (see O stelliferi conditor orbis, I:5). In many metra there is a separate closing moral (e.g. Si quantas rapidis, II:2), while in some the opening lines assume a summary function (e.g. Eheu quae miseros, III:8, in which they are treated as a refrain).

            iv) Punctuation

Punctuation provides a guide not only to understanding syntax, but also to relative degrees of hiatus between phrases.  Modern editions and primary sources can vary substantially in their punctuation.  While knowledge of both is useful for providing options for performance, the punctuation in the notated source is particularly useful as a clue to cadential placement and degrees of closure in the reconstructed melody.

            v) Voice

Awareness of voice can aid in interpreting the rhetorical style of a piece both in performance and in reconstruction. The metra in the Consolation of Philosophy are sung either by Boethius or by Philosophy, but there remain questions as to whether individual metra are sung by Boethius as a narrator (e.g. Tunc me discussa, I:3) or as an actor in the drama (most others).  There are also questions about whether Boethius and Philosophy are addressing each other, themselves as a form of internal reflection, or the Creator (e.g. O stelliferi conditor orbis, I:5, and  O qui perpetua mundi, III:9). As for affect, Boethius passes rapidly through heightened emotional states in Book I.  Philosophy is moved by Boethius’ fate at the outset, beginning with an initial cry of Heu!; thereafter she turns to initially gentle forms of persuasion and latterly denser philosophical argument.

            vi) Sound of the Text

It is particularly helpful prior to melodic reconstruction to read the poetic text aloud following prose accentuation, noting repeating patterns of word accent, especially at the end of lines.  Awareness of sounding accentual patterns can help in identifying patterns in melodic reconstruction, especially at fixed cadences.  The basic working principles of Latin pronunciation should be remembered: i) that Latin does not have an accent on the final syllable, and so two-syllable words have a stress on the first syllable; and ii) in a three-syllable word the penultimate syllable is stressed if it has a long vowel (which can be checked by consulting a dictionary), otherwise the antepenultimate syllable is stressed.  Aside from classroom scansion for the purposes of learning metre, it is most likely that late antique and early medieval poetry was read aloud with prose accent.  See, further, Dag Norberg, Les vers latins iambiques et trochaïques au Moyen Age et leurs répliques rythmiques (Uppsala, 1988), 14–16.

            vii) Metrical Design

For the metrical forms used in the De consolatione philosophiae, see the conspectus metrorum in K. Büchner (ed.), Philosophiae consolationis libri quinque: Anicius Manlius Severinus Boethius, Editiones Heidelbergensis 11 (Heidelberg: Winter), 1947 (3rd edn., 1977), 118–12.  Lupus of Ferrières (Abbot, 842-61) composed a widely disseminated early medieval guide to the metres of the De consolatione philosophiae, which parses the lyric metres into individual feet and occasionally differs from modern analyses: see Anicii Manlii Severini Boetii Philosophiae Consolationis Libri Quinque, ed. R. Peiper (Leipzig, 1871), pp. xxiiii–xxviiii. For an introduction to Latin metre, consult D. S. Raven, Latin Metre (London, 1965).  At its simplest, understanding metre is helpful in terms of locating caesuras within the line. Awareness of metre can also help in locating song models composed in similar metres, and in identifying instances of more or less systematic observance of metrical features such as ictus.


2) Read the Neumes

            i) Disposition

One of the key pieces of information provided by neumes is how stichic (i.e. linear) texts were converted into strophic songs.  A number of different strategies were used, which may be briefly summarized here:

            a) through-notation – where notation was added throughout a metrum and strophic organization was indicated through melodic repetition;

            b) notation for the opening strophe – where neumes were added only to the unit of lines used for the rest of the poem;

            c) notation for the opening strophe and occasional following lines – where neumes indicate both the routine unit of strophic return at the outset and possibilities for strophic adaptation immediately afterwards;

            d) notation for isolated internal lines – where neumes indicate a new beginning or new melodic design within a song;

Three working principles have been observed in the majority of notations: first, strophes were aligned with syntactical closure wherever possible; secondly, notators used minimal means to indicate strophic design; thirdly, simple solutions were preferred.  By way of example, a notation that opens with a four-line strophe (AABC melodically) followed by a three-line strophe (ABC melodically) signals possible strophic solutions for the remaining lines.  In the case of a metrum such as Si quantas rapidis (II:2) comprising ten couplets and syntactical closures after lines 8, 14 and 20, the most plausible solution is AABC, followed by two three-line strophes (ABC) to the end.

            ii) Interpretation

For those unfamiliar with neumatic notation, it is simplest to start by using a table that aligns neumatic signs with either square notation (as used in modern chant books) or modern equivalences.  A useful table is provided in David Hiley’s, Gregorian Chant (Cambridge, 2009), 169, which may be consulted by users in the cam domain here. A number of steps should be taken prior to reconstruction:

            a) identification of the relevant neume family;

            b) identification of individual neumes and their signalled melodic motion;

            c) identification of any further information, e.g. lines on which the neumes are aligned, use of signs that indicate the semitone step within the scale, additional letters indicating lower/higher pitches or other aspects of delivery, or neume forms indicating aspects of vocal quality (oriscus, quilisma, liquescence) – for information about several of these features, see Hiley’s Western Plainchant (Oxford, 1993), 357-61.

            d) transcription of an unheighted version of the melody – this is most simply done by recording on a blank sheet of paper the indicated melodic motion in groups of notes without stems.

            iii) Mode

Choosing a mode for realization is the most important step in reconstruction.  Consideration of a number of factors will help in guiding this decision; in particular, care should be taken to match observed features of the melody, especially intonations and cadences, with characteristic behaviours of a given mode.  A number of guides may be used in this process: beginners might find of use the explanations and tables in William P. Mahrt, ‘Gamut, Solmizaiton and Modes’, in A Performer’s Guide to Medieval Music, ed. Ross W. Duffin (Bloomington, IN, 2000), 482-495.

Matching the mood or character of the text to the mode is a necessarily more speculative enterprise.  Attempts to link modal ethos to text character have proved debatable for the Roman chant repertory, but the advice given in John’s De musica of c. 1100 on selecting modes for newly composed chants may be borne in mind for medieval song repertories.  His description of modal features is remarkable for assigning differing characters to modes on the grounds of attributed responses to specific musical features:

Alios namque morosa et curialis vagatio primi delectat, alios rauca secundi gravitas capit, alios severa et quasi indignans tertii persultatio iuvat, alios adulatorius quarti sonus attrahit, alii modesta quinti petulantia ac subitaneo ad finalem casu moventur, alii lacrimosa sexti voce mulcentur, alii mimicos septimi saltus libenter audiunt, alii decentem et quasi matronalem octavi canorem diligent.

Johannes Affligemensis, De musica cum tonario, ch. XVI, ed. Joseph Smits van Waesberghe, Corpus scriptorum de musica 1 (Rome, 1950), 109

'The slow and stately course of the first mode delights some, while the harsh gravity of the second satisfies others. The austere and somewhat indignant leap of the third pleases some, while others are attracted by the flattering sonorities of the fourth.  Some are moved by the tempered exuberance of the fifth and by its sudden descent to the final, while others are melted by the tearful voice of the sixth.  Some gladly listen to the extravagant jumps of the seventh, while others love the proper and almost matronly sonority of the eighth.'

Translation adapted from Daniel Saulnier, The Gregorian Modes, trans. Edward Schaefer (Solesmes, 1997, trans. 2002), 21.

            iv) Models

Models should be identified for reconstructed melodies wherever possible.  The most convincing models are coeval songs written in comparable verse forms. Bruno Stäblein’s edition of medieval hymn melodies provides a large body of melodies in a range of different verse forms (for references to cited studies, see below). David Hiley’s incipit index is particularly useful for navigating this volume, although what is most needed for the type of reconstruction work described here is a verse index.  Other useful resources include recent studies of neumed classical verse in the early middle ages by Silvia Wälli and Gundela Bobeth. A number of tropes were also composed in metrical verse forms: a small number of these may be found in modern editions.  Comparisons may also be made with melodic designs recorded for accentual Latin verse in this period.  Where no proximate models can be found in song forms, chants can sometimes be identified that exhibit similar behaviours.  Editions and studies that may be searched by mode greatly facilitate consultation, such as the modal index to the Graduale Romanum.

In some cases, the model that is drawn on is not a specific melody but a way of setting a text.  Some metres (especially hexameters and pentameters) are treated in the manner of a psalm recitation, while some more complex lyric metres were set by emphasizing properties of the metre. For examples, see the explanations provided in the reconstructions for Carmina qui quondam and Si quantas rapidis respectively.


Gundela Bobeth, Antike Verse in mittelalterlicher Vertonung : Neumierungen in Vergil-, Statius-, Lucan- und Terenz-Handschriften (Monumenta monodica medii aevi, Subsidia 5; Kassel and Basel, 2013)

Paul Evans, Early Trope Repertory of Saint Martial de Limoges (Princeton, N.J., 1970, repr. 2015)

David Hiley, incipit index to editions of medieval hymns

Francesco Stella and Sam Barrett (eds.), Corpus rhythmorum musicum saec. iv–ix, i: Songs in Non-liturgical Sources (Florence, 2007)

Bruno Stäblein (ed.), Hymnen: Die mittelalterlichen Hymnenmelodien des Abendlandes (Monumenta monodica medii aevi, 1; Kassel and Basel, 1956)

Sylvia Wälli, Melodien aus mittelalterlichen Horaz-Handschriften: Edition und Interpretation der Quellen (Monumenta monodica medii aevi, Subsidia 3; Kassel and Basel, 2002)