cs.CLMar 15, 2026

Echoes Across Centuries: Phonetic Signatures of Persian Poets

Authors: Kourosh Shahnazari, Seyed Moein Ayyoubzadeh, Mohammadali Keshtparvar

Organizations: Sharif University of Technology, Tehran, Iran. · Amirkabir University of Technology, Tehran, Iran.

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

This study examines phonetic texture in Persian poetry as a literary-historical phenomenon rather than a by-product of meter or a feature used only for classification. The analysis draws on a large corpus of 1,116,306 mesras from 31,988 poems written by 83 poets, restricted to five major classical meters to enable controlled comparison. Each line is converted into a grapheme-to-phoneme representation and analyzed using six phonetic metrics: hardness, sonority, sibilance, vowel ratio, phoneme entropy, and consonant-cluster ratio. Statistical models estimate poet-level differences while controlling for meter, poetic form, and line length. The results show that although meter and form explain a substantial portion of phonetic variation, they do not eliminate systematic differences between poets. Persian poetic sound therefore appears as conditioned variation within shared prosodic structures rather than as either purely individual style or simple metrical residue. A multidimensional stylistic map reveals several recurrent phonetic profiles, including high-sonority lyric styles, hardness-driven rhetorical or epic styles, sibilant mystical contours, and high-entropy complex textures. Historical analysis indicates that phonetic distributions shift across centuries, reflecting changes in genre prominence, literary institutions, and performance contexts rather than abrupt stylistic breaks. The study establishes a corpus-scale framework for phonetic analysis in Persian poetry and demonstrates how computational phonetics can contribute to literary-historical interpretation while remaining attentive to the formal structures that shape Persian verse.

Figures & tables

StageMesras
Broad mesra layer after structured extraction2,892,576
Rows with usable meter/form metadata inside the five retained classical meters1,439,014
Final inferential cohort after imposing a minimum of 2,000 mesras per poet-meter cell1,116,306
Table 1: Construction of the inferential domain from the broad corpus. Main-text models are estimated only on the final cohort.
Figure 1: Structure of the analytical corpus. The header reports the scale of the balanced cohort directly: 1,116,306 mesras, 31,988 poems, and 83 poets across the five retained classical meters. In mesra terms these meters account for 77.6% of rows with usable meter/form metadata and 100% of the final inferential cohort. Panel A shows poem counts and shares across mutaqarib , hazaj , ramal , mujtass , and muzari’ , defining the prosodic range of the controlled comparison. Panel B plots poems per century for the metadata-linked subset, showing the historical span covered by the century analysis.
Figure 2: Distributions of the three primary phonetic outcomes in the controlled analytical population. Hardness and sonority cluster around stable central tendencies, whereas sibilance has a longer upper tail, indicating that strongly strident texture is a specialized rather than ordinary condition.
Figure 3: Incremental variance explained by successive control layers in the hardness models. Meter and form account for substantial structure, but poet identity still contributes additional signal once those layers are controlled.
Figure 4: Selected controlled hardness contrasts with confidence intervals. The coefficients remain moderate in magnitude, indicating that Persian poetic style is neither metrically exhausted nor acoustically discontinuous across poets.
Figure 5: Century trajectories for hardness, sonority, sibilance, and entropy with poem-bootstrap uncertainty bands. The curves show redistribution across the attested centuries of the corpus rather than a single linear shift from “hard” to “soft.”
Figure 6: Phonetic stylistic space of Persian poets. Grey points show the wider field; colored points mark the focal case studies and historical anchors, with the cluster legend placed outside the plotting area.
ArchetypeMetric signatureRepresentative poetsLiterary interpretation
High-sonority lyricHigh sonority and vowel ratio, low compaction, usually low-to-moderate hardnessShah Nematollah Vali, Fakhr Al-Din Asad Gorgani, Jahan Malek Khatun, Rumi, Shahriar (descriptive)Open vocalic texture, sustained lyric or contemplative cadence, and reduced consonantal congestion without acoustic blandness.
High-hardness rhetoricalHigh hardness with cluster support and restrained sibilanceSaadi, Ferdowsi, Asadi Tusi, Khayyam (descriptive)Aphoristic, epic, or didactic firmness in which authority comes from consonantal closure more than from fricative brightness.
Sibilant mystical / rhetoricalElevated sibilance with moderate hardness and strong local articulationSafi Alishah, Khaqani, Hafez, BolandeghbalCourtly, devotional, or lyric sharpness produced by contour and strident edge rather than by maximal pressure.
High-entropy complexityHigh entropy with mixed hardness and sonority, often with cluster supportBidel Dehlavi, Ashofteh Shirazi, Saeb, Juya TabriziLater Persianate or return-style texture in which symbolic variety and local contrast become the primary vehicles of distinctiveness.
Table 2: Phonetic archetypes in the Persian stylistic space. Archetypes are descriptive regions rather than discrete school labels, and they summarize the stylistic map discussed in the main text rather than imposing a rigid taxonomy on poets or periods.
Code / meterScansionRetained sub-patternPoemsMesrasHard.Son.Sib.Ent.
M01 / Mutaqaribfa’ulun fa’ulun fa’ulun fa’ulmutaqarib-i musamman-i mahzuf3,619338,2501.6083.1930.1793.707
M02 / Hazajmafa’ilun mafa’ilun fa’ulunhazaj-i musaddas-i mahzuf4,749236,0941.5823.2170.1833.715
M03 / Ramalfa’ilatun fa’ilatun fa’ilatun fa’ilunramal-i musamman-i mahzuf9,690198,7261.5793.2160.1763.873
M04 / Mujtassmafa’ilun fa’ilatun mafa’ilun fa’ilunmujtass-i musamman-i makhbun-i mahzuf7,334193,0741.6013.2470.1873.817
M05 / Muzari’maf’ulu fa’ilatu mafa’ilu fa’ilunmuzari’-i musamman-i akhrab-i makfuf-i mahzuf6,603150,6521.6033.2210.1863.818
Table 3: Most common prosodic meters in the balanced analytical cohort, with poem-level phonetic means. Together the five retained meters account for 77.6% of mesras with usable meter/form metadata and 100% of the final controlled cohort; within that cohort their mesra shares are 30.3% ( mutaqarib ), 21.1% ( hazaj ), 17.8% ( ramal ), 17.3% ( mujtass ), and 13.5% ( muzari’ ). Internal corpus labels are mapped to standard Persian names for the relevant bahrs and to the exact retained sub-patterns so that the computational codes remain readable in literary-historical terms. Metrics are averaged within poems before meter-level aggregation so that long works do not dominate the comparison.
MeterGenre / historical associationRecurrent phonetic affordanceIllustrative poets
MutaqaribEpic masnavi , heroic narration, and later didactic-public reuse; especially prominent in early classical epic memoryRepeated cadential closure and compact consonantal sequencing support firmness and cluster-backed assertion more readily than in lyric-dominant meters.Ferdowsi, Saadi, Bolandeghbal
HazajRomance masnavi , chantable lyric extension, and later devotional or pedagogical reuseFavors smoother linear motion and higher sonority, yet can still carry sibilant incision when devotional diction sharpens the line.Fakhr Al-Din Asad Gorgani, Safi Alishah, Jahan Malek Khatun
RamalGhazal, mystical lyric, and reflective didactic verse; especially central from the high-classical lyric centuries onwardPreserves strong contrast on the hardness-sonority axis rather than forcing a single texture; contemplative softness and pressure-rich intricacy can coexist within the same envelope.Shah Nematollah Vali, Bidel Dehlavi, Khaqani
MujtassCompact ghazal and courtly lyric, particularly in performance-oriented lyric repertoiresEncourages openness, vowel prominence, and local contour, making it hospitable to bright lyric memorability without excessive compaction.Hafez, Jahan Malek Khatun, Ashofteh Shirazi
Muzari’Rhetorical lyric and learned late-classical, Indo-Persian, and revivalist texturesHeightens trade-offs between hardness, contour, and entropy; poets can remain firmness-forward, contour-forward, or diversity-rich while sharing the same metrical frame.Khaqani, Saadi, Bidel Dehlavi, Ashofteh Shirazi
Table 4: Meter interaction as phonetic affordance. The table summarizes corpus-supported genre and period associations together with recurrent sonic possibilities, not fixed meter essences.
Figure 7: Case-study fingerprint atlas for thirteen focal poets and comparative anchors. Every panel uses the same metric order and the same 0–1 radial normalization derived from the full poet atlas, so differences can be compared directly across poets rather than inferred from separate scales. Ferdowsi, Saadi, and Khayyam lean toward hardness and closure; Hafez and Jahan Malek Khatun toward sonority and vowel spread; Safi Alishah and Khaqani toward sibilant contour; Bidel toward entropy and cluster pressure; and Shahriar toward modern descriptive openness.
PoetMutaqaribHazajRamalMujtassMuzari’
Hafez–––0.03/0.17/-0.03–
Khaqani––0.04/-0.00/0.00–0.07/-0.06/0.09
Saadi0.25/-0.11/-0.01––0.10/0.03/-0.240.21/-0.12/-0.09
Bidel Dehlavi––0.22/-0.15/0.060.13/-0.06/0.020.15/-0.24/0.06
Safi Alishah–0.06/-0.20/0.08–––
Shah Nematollah Vali––-0.35/0.26/-0.10––
Table 5: Case-study meter invariance summary for the balanced inferential subset only. Each cell reports standardized within-meter effects for Hardness/Sonority/Sibilance; dashes indicate poet-meter cells without sufficient mesra support. Columns correspond to the retained named meters mutaqarib , hazaj , ramal , mujtass , and muzari’ . Descriptive anchors such as Ferdowsi, Shahriar, and Khayyam are discussed in the text but do not appear here because the table is restricted to supported within-meter estimates.
Figure 8: Within-meter comparison for the supported case-study subset. Columns correspond to the retained classical meters mutaqarib , hazaj , ramal , mujtass , and muzari’ , and rows to focal poets with sufficient mesra support in at least one cell. The three aligned heatmaps report standardized within-meter effects for hardness, sonority, and sibilance; red cells indicate positive deviations, blue cells negative deviations, and gray cells unsupported poet-meter combinations. The figure highlights Saadi’s portable hardness, Bidel’s multiaxial contrast across several meters, Shah Nematollah Vali’s strong ramal reversal toward low hardness and high sonority, and the more localized contour patterns of Hafez and Safi Alishah.
Appendix figures & tables11 assets

Supplementary material from the paper’s appendix.

Appendix

PoetMesras
Attar109514
Ferdowsi99220
Saeb Tabrizi96672
Iranshan41650
Khwaju Kermani41366
Nezami33166
Appendix
Table 14
Figure 9: Appendix atlas page 1, covering Attar through Qaaani Shirazi. The fixed radial scale allows direct comparison of openness, firmness, contour, entropy, and cluster pressure across poets from different periods and literary registers.
Figure 10: Appendix atlas page 2, covering Anonymous Faramarznama through Ibn Yamin. The fixed radial scale allows direct comparison of openness, firmness, contour, entropy, and cluster pressure across poets from different periods and literary registers.
Figure 11: Appendix atlas page 3, covering Salim Tehrani through Naziri Nishaburi. The fixed radial scale allows direct comparison of openness, firmness, contour, entropy, and cluster pressure across poets from different periods and literary registers.
Figure 12: Appendix atlas page 4, covering Saif Farghani through Hazin Lahiji. The fixed radial scale allows direct comparison of openness, firmness, contour, entropy, and cluster pressure across poets from different periods and literary registers.
Figure 13: Appendix atlas page 5, covering Bolandeghbal through Saghir Isfahani. The fixed radial scale allows direct comparison of openness, firmness, contour, entropy, and cluster pressure across poets from different periods and literary registers.
Figure 14: Appendix atlas page 6, covering Mirza Aqa Khan Kermani through Ohadi Maraghei. The fixed radial scale allows direct comparison of openness, firmness, contour, entropy, and cluster pressure across poets from different periods and literary registers.
Figure 15: Appendix atlas page 7, covering Toghral Ahrari through Shah Nematollah Vali. The fixed radial scale allows direct comparison of openness, firmness, contour, entropy, and cluster pressure across poets from different periods and literary registers.
PoetMutaqaribHazajRamalMujtassMuzari’
Adib Al-Mamalek Farahani0.000.000.000.00–
Adib Saber–––-0.150.00
Afsar Al-Moluk Ameli-0.06––––
Ahli Shirazi–-0.050.090.100.15
Amir Khusrow Dehlavi–0.150.110.080.12
Amir Moezzi––0.080.040.18
Appendix
Table 22
PoetMutaqaribHazajRamalMujtassMuzari’
Adib Al-Mamalek Farahani0.000.000.000.00–
Adib Saber–––0.200.00
Afsar Al-Moluk Ameli-0.06––––
Ahli Shirazi–-0.02-0.010.02-0.18
Amir Khusrow Dehlavi–0.110.150.11-0.06
Amir Moezzi––0.060.09-0.11
Appendix
Table 23
PoetMutaqaribHazajRamalMujtassMuzari’
Adib Al-Mamalek Farahani0.000.000.000.00–
Adib Saber–––0.130.00
Afsar Al-Moluk Ameli-0.06––––
Ahli Shirazi–0.01-0.04-0.10-0.05
Amir Khusrow Dehlavi–-0.18-0.15-0.13-0.08
Amir Moezzi––-0.08-0.010.04
Appendix
Table 24

Explore similar work

CardsList
Apr 8, 2026

Between Century and Poet: Graph-Based Lexical Semantic Change in Persian Poetry

Meaning in Persian poetry is both historical and relational. Words persist through literary tradition while shifting their force through changing constellations of neighbors, rhetorical frames, and poetic voices. This study examines that process using aligned Word2Vec spaces combined with graph-based neighborhood analysis across centuries and major poets. Rather than modeling semantic change as vector displacement alone, it treats lexical history as the rewiring of local semantic graphs: the gain and loss of neighbors, shifts in bridge roles, and movement across communities. The analysis centers on twenty target words, anchored by five recurrent reference terms: Earth, Night, two wine terms, and Heart. Surrounding them are affective, courtly, elemental, and Sufi concepts such as Love, Sorrow, Dervish, King, Annihilation, and Truth. These words exhibit distinct patterns of change. Night is more time-sensitive, Earth more poet-sensitive, and Heart shows continuity despite graph-role mobility. The two wine terms highlight probe sensitivity: one is broad and semantically diffuse, while the other is narrower and more stable. A lexical audit confirms that the corpus contains historically driven terms, poet-specific usages, and sparsely attested mystical vocabulary requiring caution. Overall, semantic change in Persian poetry is better captured as neighborhood rewiring than as abstract drift. For Digital Humanities, this approach restores local structure to computational analysis and supports interpretations closer to literary practice: persistence, migration, mediation, and selective transformation.

Kourosh Shahnazari, Seyed Moein Ayyoubzadeh, Mohammadali Keshtparvar

Sharif University of Technology · Amirkabir University of Technology

Semantic Change DetectionComputational Literary Studies
AskPDF
Apr 1, 2026

A Dynamic Atlas of Persian Poetic Symbolism: Families, Fields, and the Historical Rewiring of Meaning

Persian poetry is often remembered through recurrent symbols before it is remembered through plot. Wine vessels, gardens, flames, sacred titles, bodily beauty, and courtly names return across centuries, yet computational work still tends to flatten this material into isolated words or broad document semantics. That misses a practical unit of organization in Persian poetics: related forms travel as families and gain force through recurring relations. Using a corpus of 129,451 poems, we consolidate recurrent forms into traceable families, separate imagistic material from sacred and courtly reference, and map their relations in a multi-layer graph. The symbolic core is relatively sparse, the referential component much denser, and the attachment zone between them selective rather than diffuse. Across 11 Hijri-century bins, some families remain widely distributed, especially Shab (Night), Ruz (Day), and Khaak (Earth). Wine vessels, garden space, flame, and lyric sound strengthen later, while prestige-coded and heroic-courtly vocabulary is weighted earlier. Century-specific graphs show change in arrangement as well as membership. Modularity rises, cross-scope linkage declines, courtly bridges weaken, and sacred bridges strengthen. Hub positions shift too: Kherqe (Sufi Robe) gains late prominence, Farkhondeh {Blessed} and Banafsheh (Violet) recede, and Saaghar (Wine Cup) stays central across the chronology. In this corpus, Persian symbolism appears less as a fixed repertory than as a long-lived system whose internal weights and connections change over time.

Kourosh Shahnazari, Seyed Moein Ayyoubzadeh, Mohammadali Keshtparvar

Sharif University of Technology, Tehran, Iran. · Amirkabir University of Technology, Tehran, Iran.

Semantic Change DetectionComputational Literary Studies
AskPDF
Mar 11, 2026

Artificial Intelligence for Sentiment Analysis of Persian Poetry

Recent advancements of the Artificial Intelligence (AI) have led to the development of large language models (LLMs) that are capable of understanding, analysing, and creating textual data. These language models open a significant opportunity in analyzing the literature and more specifically poetry. In the present work, we employ multiple Bidirectional encoder representations from transformers (BERT) and Generative Pre-trained Transformer (GPT) based language models to analyze the works of two prominent Persian poets: Jalal al-Din Muhammad Rumi (Rumi) and Parvin E'tesami. The main objective of this research is to investigate the capability of the modern language models in grasping complexities of the Persian poetry and explore potential correlations between the poems' sentiment and their meters. Our findings in this study indicates that GPT4o language model can reliably be used in analysis of Persian poetry. Furthermore, the results of our sentiment analysis revealed that in general, Rumi's poems express happier sentiments compared to Parvin E'tesami's poems. Furthermore, comparing the utilization of poetic meters highlighted Rumi's poems superiority in using meters to express a wider variety of sentiments. These findings are significant as they confirm that LLMs can be effectively applied in conducting computer-based semantic studies, where human interpretations are not required, and thereby significantly reducing potential biases in the analysis.

Arash Zargar, Abolfazl Moshiri, Mitra Shafaei +3

Department of Mechanical & Industrial Engineering; Vector Institute for Artificial Intelligence; University of Toronto; Toronto, Canada · Department of Near & Middle Eastern Civilizations; University of Toronto; Toronto, Canada · Independent Researcher +1

Computational Literary StudiesSentiment Analysis
AskPDF