Submissions

Login or Register to make a submission.

Author Guidelines

General Information

Since Volume 51 (March 2025), Computational Linguistics has adopted a new vision. Under its new vision, the journal strongly encourages research that engages with emerging paradigms, with particular emphasis on contemporary language model technologies (e.g., Large Language Models, or LLMs) and other transformative developments reshaping computational language research. At the same time, the journal remains committed to insightful contributions grounded in theoretical, mathematical, algorithmic, linguistic, and interdisciplinary perspectives. We particularly welcome studies that bridge these foundational principles with the latest scientific advancements. Our mission is for Computational Linguistics to serve as a central forum where diverse perspectives converge and collaborate to drive the field forward. Please refer to this Editorial for more information.

Manuscript format
  • A single-spaced PDF formatted according to our Style Guidelines, including an abstract.
  • Written in English.
  • When submitting, select the article type as the journal section, or the name of the special issue if you are submitting to one (in that case, state the paper type in your cover letter).
  • Computational Linguistics does not charge any submission, processing or publication fees.

Substantially extended versions of conference papers may be acceptable as submissions. In such cases, the manuscript must clearly explain how the work extends beyond the earlier publication and why the contributions of the extended version merit presentation in this central venue and deserve the community's attention. This will help both reviewers and readers readily understand the value of the work.

Is CL the right venue?
  • Scope. We publish work that makes a substantive contribution to the computational processing of language, generally from a natural language processing perspective. A good diagnostic is whether a significant proportion of your references are to publications in this area (see the journals, conferences and workshops listed on the ACL Wiki). If not, the paper is probably better suited to another venue.
  • Conference-length work. A 6–8 page double-column ACL conference paper (about 10–12 pages in the journal style) typically does not contain sufficient content for a CL journal article. Such work is usually best submitted to a conference or workshop first, and later developed into a more substantial report.

Article Types at a Glance

Type Length / how to submit
Long Papers Typically up to 40 pages of main text
Short Papers Typically up to 20 pages (final version up to 25)
Survey Articles Submit a proposal first (outline up to 5 pages)
Position Papers Well-argued perspectives; experiments not required
Squibs and Discussions Up to 8 pages of content, plus references
Last Words Up to 10 pages; contact the Editor-in-Chief first
Book Reviews Contact the Book Review Editor
Award Articles By arrangement with the editorial office

Page lengths refer to the main content of the paper (including the title, authors, affiliations and abstract) and exclude acknowledgements, references and appendices.

Long Papers

Long Papers present comprehensive and substantial research contributions in computational linguistics and natural language processing. These may include novel theoretical models, extensive empirical analyses, or large-scale system designs. Papers in this category typically involve deep and broad exploration, with significant methodological or conceptual innovation. All submissions are held to the highest standards of technical depth, clarity, and impact.

Page length: typically up to 40 pages of main text.

Short Papers

Short Papers report focused, high-quality research findings that may offer innovative ideas, efficient methods, or surprising results that advance the field. This track is ideal for compact contributions with strong insights. Well-analyzed negative results or innovative applications are also suitable for this track. Like long papers, short papers are reviewed rigorously and held to the same high standards of scientific and presentation quality.

Page length: initial submissions are typically up to 20 pages. Final versions may be up to 25 pages if needed to incorporate revisions or improve clarity.

Survey Articles (and Proposals)

We welcome survey articles that offer comprehensive, insightful overviews of specific subfields or emerging interdisciplinary areas within computational linguistics and natural language processing. These articles serve as both valuable entry points for newcomers and authoritative references for experienced researchers.

We particularly encourage two types of surveys:

  • Core Area Surveys: These articles provide an in-depth synthesis of the state of the art in well-established subfields of CL/NLP, consolidating diverse research threads into a coherent narrative. Such surveys are ideal for helping researchers stay current outside their immediate specializations and can serve as foundational resources for graduate instruction.
  • Peripheral and Interdisciplinary Surveys: These explore research areas at the interface of CL/NLP and adjacent disciplines (e.g., cognitive science, social science, multimodal processing) that are currently underrepresented in the journal. The goal is to broaden the journal’s scope and introduce new perspectives to the community.

A high-quality survey article should not merely catalog existing work, but should provide critical interpretation, identify trends, open problems, and opportunities for future research. Where relevant, surveys may complement existing textbook material by offering significantly deeper or more up-to-date coverage. There is no fixed limit on the number of survey articles published each year.

Submit a survey proposal first. The proposal should be a single PDF file formatted using the latest CL style file, and it must include:
  • a title and the names/affiliations of the authors;
  • an outline (excluding references) of up to 5 pages detailing the content of the proposed article;
  • short (one-paragraph) biographies of the authors;
  • the names of three recognized experts qualified to review the submission, with no conflict of interest with the authors.
The outline must succinctly describe the content of the survey and clearly identify the body of work the article will cover by including numerous relevant citations and a bibliography. Proposals will be reviewed by the editorial board through an accelerated review process.

Position Papers

Position papers play a crucial role in advancing scholarly discourse by allowing researchers to present well-argued perspectives on key challenges, unresolved issues, and emerging opportunities in computational linguistics and natural language processing.

While not necessarily requiring experimental results, position papers published in Computational Linguistics must demonstrate rigorous reasoning and be supported by persuasive evidence to substantiate their stance on major topics or future directions in the field.

Squibs and Discussions

This category is reserved for very short articles that constitute more than programmatic versions of regular papers. Squibs should possess at least one of the following attributes:

  • unexpectedness, e.g., a demonstration that a commonly accepted idea or method is flawed;
  • genuine novelty, e.g., thus-far unnoticed language data that challenges current methods; and
  • being targeted to a large segment of our readership.

Articles submitted as squibs that are deemed to be merely insufficiently developed short papers will be desk rejected. Papers about language resources may be acceptable provided the relevant resources are truly novel and of general interest.

Page length: must not exceed 8 pages of content (with unlimited pages for references).

Last Words

In each issue of the journal we reserve a small number of pages for a personal opinion or provocative perspective on some aspect of the field of computational linguistics. You can see all previous Last Words pieces here. Contact the Editor-in-Chief directly if you are interested in submitting such a piece.

Page length: up to 10 pages.

Book Reviews

Anyone interested in reviewing a book, or in suggesting a book for review, should contact the Book Review Editor:

Aline Villavicencio
University of Exeter and University of Sheffield
a.villavicencio@sheffield.ac.uk

Award Articles

Award articles, including ACL Lifetime Achievement Award articles and ACL Computational Linguistics Doctoral Dissertation Award articles, will be published directly through the editorial office by separate arrangement with the award recipients.

Submission Preparation Checklist

Submission Declaration. By submitting, the corresponding author confirms on behalf of all authors that the submission complies with the ACL Policy on Publication Ethics, and in particular that:

  1. Dual submission. The manuscript has not been published in an archival venue and is not under consideration elsewhere (including journals, conferences, and ACL Rolling Review). This includes shorter or other versions of this work that are under review at, or planned for submission to, archival venues while this manuscript is under consideration. Substantially extended versions of published conference papers are acceptable (see item 2). Manuscripts previously declined by Computational Linguistics will not be considered.
  2. Prior publications and text re-use. All of the authors' prior or concurrent work similar to or overlapping with this manuscript (published or under review) is cited in the manuscript, which clearly states how this submission differs (e.g., in the introduction). These works are also listed in the cover letter; the editors may request copies of unpublished ones. Preprints (e.g., on arXiv) and non-archival presentations of this work need not be listed. Any material re-used from earlier work is clearly indicated and referenced.
  3. Authorship. All authors meet the ACL authorship criteria and have approved the submission. All authors are entered in the submission system, with full names, affiliations, and email addresses matching the manuscript. Affiliations are stated accurately, and institutional email addresses are used whenever possible rather than generic ones (e.g., Gmail).
  4. Generative AI. Generative AI tools are not listed as authors. Any use of generative AI to create content (e.g., text, code, analyses, data, or research ideas) is disclosed in the Acknowledgements section. Grammar or spell checkers and predictive typing need not be disclosed. The authors are responsible for all content, including the accuracy of every citation. The manuscript contains no hidden or obscured prompts intended to manipulate its automated processing.
  5. Research ethics. Research involving human participants has IRB (or equivalent) approval and informed consent. Where no such body exists, the manuscript justifies the safeguards adopted.
  6. Data, licenses, and permissions. Data, tools, and materials are used in accordance with their licenses and terms of use, and any required permissions were obtained before submission. The manuscript states the availability of data and code. Anything stated as available will be released by the time of publication.
  7. Competing interests. All relevant competing interests, including financial relationships, patents, and commercial products or services, are disclosed in the submission form.
  8. Completeness. All supplementary materials referred to in the manuscript are included or accessible via the links provided.
  9. Coordinated disclosure. Work reporting new security weaknesses or adversarial techniques follows the ACL coordinated-disclosure requirements; the disclosure process is documented and available to the editors on request.
  10. Suggested experts (survey proposals only). The three experts named in a survey proposal have no conflict of interest with the authors and are listed with their institutional email addresses.
  11. Format. The manuscript is a single-spaced PDF prepared with the current CL style files (without the manuscript option), with author names and affiliations on the first page (CL uses single-blind review).
  12. Disclosure agreement. The submission is not covered by a non-disclosure agreement and is available for peer review without restriction.
  13. Copyright. The authors are able to assign copyright to the Association for Computational Linguistics, should the submission be accepted.

Cover letter. A cover letter is uploaded as a separate file (file type "Cover.Letter") and includes:

  1. the submission type and a brief statement of the main contributions;
  2. 3–6 keywords describing the work (also entered in the keywords field of the submission form);
  3. the potential impact of the work on the CL/NLP community and beyond, and why Computational Linguistics, as the central forum of the CL/NLP community, is the right venue for it;
  4. the list of prior or concurrent work similar to or overlapping with this manuscript, with how this submission differs, and, for an extended conference paper, what is new (item 2).

Comments for the Editor. Use this field in the submission form for information intended for the editors only and not to be shared with reviewers: in particular, if the manuscript was previously declined elsewhere (e.g., at an *ACL conference, ACL Rolling Review, another journal, or a venue such as ICLR): a summary of the changes made in response to the earlier reviews, the reviews received and a link to the previous version (or a copy on request). Otherwise, it may be left empty.

Detailed instructions: original submissions; resubmissions and final versions.

Privacy Statement

The names, email addresses, and other information entered in this journal site will be used exclusively for the stated purposes of this journal and will not be made available for any other purpose or to any other party, except as described below or as required to support the journal's editorial and publishing operations.

To uphold the integrity and quality of the peer review and publication process, Computational Linguistics may, at any stage of the editorial workflow, submit manuscripts and associated materials to trusted third-party screening services, such as the similarity-detection service iThenticate and other automated screening tools. Such tools are used strictly for editorial purposes and under conditions that protect the confidentiality of submissions; in particular, these services are used under settings that ensure submitted content is processed solely for performing the requested checks, is not used to train AI models, and is not retained beyond a limited period, except as required by law or to prevent misuse. All editorial decisions remain the responsibility of the journal's editors. By submitting to the journal, authors acknowledge and consent to such processing.