h t t p s : / / w w w . s t a t . c m u . e d u / ~ b r i a n / z a c h - 0 0 / 7 6 5 - 2 0 1 9 / 7 6 4 - 2 0 1 6 / w e e k 0 5 / r e f e r e e i n g / L a r r y - A b o l i s h - t h e - P e e r - R e v i e w . p d f

Ingest: A World Without Referees

Type: types/ingest-report.md

Classification

A polemical conceptual essay proposing institutional change, dated February 20, 2012. It argues from professional experience, analogies, and an alternative publishing arrangement rather than a comparative study. Author: Larry Wasserman, Carnegie Mellon University; he reports 24 years of experience as an author, referee, associate editor, and editor.

Summary

Wasserman argues that journal refereeing delays dissemination and gives a few fallible reviewers excessive control over which research becomes available. He proposes posting papers on arXiv or personal websites, inviting criticism, and revising when readers find errors. A proposal he quotes from Walter Noll adds public reviews and discussions, author notification of colleagues, and promotion credit for reviewing. Discovery and judgment would move to readers and voluntary review services. Wasserman retains refereeing for grants and principally addresses familiar fields such as statistics and computer science. His essay makes a case for separating publication from certification, but does not test whether voluntary scrutiny supplies adequate coverage or improves research quality.

Quotes

No source quotes have been retained yet.

Connections Found

The essay is a policy comparison for the separation between candidate retention and reliance: making an idea available for criticism does not establish what uses its support licenses. Wasserman argues for removing a publication barrier; Commonplace must also decide when retained material may guide an agent. This distinction limits the apparent conflict with false-positive acceptance becoming operative. Open publication moves assessment to readers and later critics, while operative acceptance concerns a claim actually shaping behavior. The essay does not establish that these later assessments are reliable enough for agent consumption.

Extractable Value

  1. Separate permission to circulate from permission to rely. The publishing proposal offers a concrete institutional comparison for the KB's existing candidate-versus-reliance distinction. A retention policy should state which boundary a review gate controls; removing a barrier to circulation need not remove assessment before use. This is a conceptual illustration, not evidence that either policy improves outcomes. [quick-win]
  2. Treat criticism and discovery as services with their own incentives. Noll's quoted proposal explicitly recognizes reviews in promotion, while Wasserman expects independent discovery services to emerge. The combination identifies work that survives removal of journal admission: finding relevant material, criticizing it, and responding to criticism. Whether those incentives provide enough attention remains an empirical question. [just-a-reference]

Limitations (our opinion)

Accounts of unfair rejection and reviewer disagreement motivate concern but do not measure its prevalence or the net benefit of review. Fallible referees can still improve papers or selection on average. The simpler explanation that a useful institution makes errors remains compatible with the reported experiences, so they do not establish that abolition is better than reform.

The proposed replacement assumes that good ideas attract attention, bad ideas are ignored or corrected, and voluntary discovery services cover readers' needs. No comparison measures these outcomes, the distribution of attention, or correction delays. The appeal to arXiv use in mathematics and physics does not isolate a world without journal refereeing from a world where preprints and journals coexist. The wine-market analogy establishes neither the costs of erroneous scientific reliance nor the reader effort needed to assess papers.

The author acknowledges medical risk but offers no developed account of it and narrows his main focus to fields he knows. Transfer to an agent-operated KB is likewise limited: source availability is not equivalent to automatic adoption. As the false-positive acceptance note explains, correction after use still consumes evaluation capacity, and the accepted claim may have affected behavior before correction. The essay's historical account of refereeing is not independently established here.

File this ingest as a source-only reference for distinguishing dissemination from warranted reliance; the existing candidate-retention note already supplies the transferable distinction, and this essay adds no tested basis for changing Commonplace's review policy.