Talk2NatureParrots / Vocal variation and social networks
Parrot social ties track call variation, not call meaning
Published May 7, 2025 · Note reviewed 2026-10-23
This study combines social-network observations with monk-parakeet vocalizations. It reports networks for 337 birds, relatedness for 100, and 5,599 vocalizations from 229 individuals over two years. Call-diversity associations varied by year; close affiliates had more dissimilar calls. These are observational associations, not evidence that a specific call carries a known message.
What we reviewed · Targeted article, deposit-metadata and pinned-repository review; metadata only
Reviewed the published article abstract and targeted methods, results, discussion, data-access, ethics and rights passages using the primary full-text record and institutional copy; checked the 2023 voiceprint paper for sample overlap; reviewed the Zenodo record and the GitHub README/tree at a pinned revision. On October 24, rechecked Methods section 2.1 and confirmed the Barcelona study region and field dates for the atlas. The archive, raw recordings, code and model outputs were not acquired or run.
Not yet independently reviewed by a parrot-bioacoustics, animal-behavior or machine-learning specialist.
Study at a glance
- Question
- How do social-network position and relationships relate to individual vocal diversity and contact-call similarity?
- Study sample
- 337 monk parakeets in social-network observations, relatedness for 100, and 5,599 vocalizations from 229 individually marked birds over two years in Barcelona.
- Reported patterns
- The paper reports year-specific relationships between call diversity, age and nest-mate counts, and more dissimilar calls among strongly affiliated birds. These patterns are associations, not causal estimates of what calls mean.
- Overlap with prior work
- The 2023 voiceprint paper reports the same 5,599 calls from 229 birds over two years. Treat this 2025 analysis as an extension of that cohort, not an independent replication.
- Asset status
- The article declares CC BY 4.0; Zenodo v1 declares MIT for its deposit; the pinned GitHub tree has no located license. No archive, code or audio was acquired or executed.
Reading location: Smeele et al. (2025), abstract and targeted Methods, Results, Discussion, Data accessibility, Ethics and copyright passages; Smeele et al. (2023), Results and study-sample description; Zenodo record 10.5281/zenodo.13906800; GitHub commit 395a622b18f3f55fd4efdcb9a093cc214f6fdf67.
What this does — and doesn’t — establish
The exact 5,599-call and 229-bird totals match the 2023 monk-parakeet voiceprint paper, so this is a new analysis of an overlapping cohort, not an independent acoustic replication. The study does not test receiver responses or experimentally established meanings. The authors discuss active differentiation as an interpretation; the reported associations alone do not establish that cause. The Zenodo archive was not inspected, and the GitHub README says some required inputs are not shared. Its documented nest/UTM location fields make location review necessary before any mapping. No coordinates are reproduced here.
What we can build on
This work shows why social relationships and sound should be recorded as separate, linked evidence. A useful next study could test whether carefully observed social context improves a bounded caller-identity or call-variation task, with whole-bird and whole-session holdouts. That would still be distinct from decoding message content or proving that a receiver understood a call.
A practical starting point
Metadata-only; not admitted to Talk2Nature training or public app inference.
Available starting point: Article metadata and targeted text review, Zenodo record metadata and pinned GitHub README/tree only. No audio, dataset archive, code or model was acquired or run.
Reuse limits: The article declares CC BY 4.0; the Zenodo deposit declares MIT at record level, but its member files and embedded-media provenance were not inspected; no GitHub repository license was located at the pinned revision. Keep these assets separate and resolve component-specific terms before use.
Our proposed experiment: No experiment or model was run. A future analysis would need a verified call/social-event join, overlap and missingness audit, independent specialist review, rights and location review, and animal/session-grouped evaluation. Social association does not establish call meaning or receiver comprehension.
See the related tools and reading paths →
The next reading task
Ask an independent parrot-bioacoustics or social-network specialist to judge whether the paper’s network measures can be joined to call-level events and what missingness would mean. Before archive acquisition, verify the exact overlap with the 2023 release, inspect the manifest and member-level terms, and review the location fields and missing inputs. Keep the 2025 code metadata-only until a license and dependencies are reviewed; do not publish fine-grained locations.
Go to the sources
- The effect of social structure on vocal flexibility in monk parakeets (2025) ↗
Smeele et al., Royal Society Open Science 12(5):241717. The published study relates social-network measures to vocal diversity and contact-call similarity in wild monk parakeets.
- Monk-parakeet social-structure data deposit, Zenodo v1 ↗
Official Zenodo record 10.5281/zenodo.13906800 lists one 129.5 MB ZIP and declares the MIT License at record level.
- Pinned analysis repository for monk-parakeet social structure (commit 395a622) ↗
Author GitHub repository at immutable commit 395a622b18f3f55fd4efdcb9a093cc214f6fdf67. Its README describes R/Stan analyses and notes that some input data are not shared.
- Monk parakeet vocal signatures ↗
Smeele et al. (2023), final Royal Society Open Science article. Targeted abstract and methods/figure text reviewed from the primary PMC index; direct full-text retrieval was CAPTCHA-limited and the ANU repository PDF fetch timed out. The ANU record lists Open Access under CC BY 4.0.
Original editorial note by Talk2Nature contributors. AI-assisted, source-grounded synthesis; not a peer review or an independent replication. Linked works retain their own rights.