The Ladder of Abstraction: Understanding Communication Gaps
The ladder of abstraction reveals why vague language can create barriers in effective communication.
What was corrected
The article stated: 'Published data from the project, which Google released in 2015, found that the highest-performing teams were distinguished by a communication norm that the researchers coded as concrete contributions before abstract principles', naming Abeer Dubey, Julia Rozovsky and Anita Woolley, and then added in the same passage that Project Aristotle 'did not report a concrete-versus-abstract communication coding scheme or a decision-quality effect size of this kind'. It described Eagly and Chaiken running 'a series of experiments' whose audiences persuaded by concrete arguments 'showed attitude change that was significantly more resistant to counterargument one week later'. It said CLT 'has been tested across over 200 studies'. It headed a section 'Steven Pinker's corpus analysis of effective prose' while conceding he 'presents this as a stylistic observation rather than a quantified corpus study'. It claimed EQC 'contracted Plain Language Group New Zealand', that the revision 'reduced inbound telephone queries by 4substantially within about 18 months', and that the government 'cited the EQC reform as a model' in the Plain Language Act 2022. It said Geisler 'studied how experts and novices in philosophy and medicine worked with the same written texts' and that experts 'have lost the capacity to inhabit the lower levels they once occupied'. It said 'The high-abstraction language allowed the crisis to grow substantially', that mortgages went 'to people who could not afford them' and were sold to institutions 'that did not understand what they had bought', that Enron's abstraction layer 'provided cover', that Intel's stock 'fell 9% in the week following Nicely's concrete disclosure' and that Intel 'established a new industry standard', and that the Columbia board 'documented a systematic abstraction mismatch'. Mission problems 'almost always involve ladder mismatch'; 'what specifically do we mean' was 'the most powerful question'; floating abstractions were 'covers for the absence of thinking'.
The Project Aristotle section was deleted entirely rather than left standing with a trailing correction, as was the Eagly and Chaiken programme, the over-200-studies figure, and the whole Earthquake Commission case including the corrupted 'by 4substantially' statistic. The Pinker section is now described as stylistic observation, and the example-principle-application pattern is labelled a WhenNotesFly framework. Geisler is correctly described as examining reading, writing and knowing in academic philosophy, and the lost-capacity claim is replaced with the point that experts may find it harder to reconstruct a novice reading because their knowledge is organised differently, while still being able to explain concretely and test with real readers. Construal Level Theory is presented as adjacent evidence rather than validation of Hayakawa's model, with Liberman and Trope 1998 and Trope and Liberman 2010 cited directly and the matching effect described as varying by task, goal and audience. The financial crisis passage now says abstract terminology may have made the crisis harder for non-specialists to understand rather than allowing it to grow, and describes poorly underwritten mortgages packaged into securities through a system that underestimated correlated default risk. Enron is framed as illustrating the gap between abstract values and concrete conduct without claiming the values caused the fraud. Columbia was removed, leaving one Intel Pentium case that restores the replacement policy, credibility, acceptable-error and IBM factors and drops the unsourced stock figure and industry-standard claim. The one-floor-above heuristic and the Bessie table are labelled WNF adaptations, jargon is separated from familiarity and from abstract generality with concrete technical terms as counterexamples, and a new passage states that high abstraction is essential for theory, law, strategy and transfer so the model reads as fit between level and purpose. Universal wording was replaced throughout, the FAQ no longer prescribes a single sequence, and the article went from about 4,790 to about 2,550 words.
Suggested change
Deleted the fabricated Project Aristotle finding in full rather than leaving it with a trailing disclaimer, and removed the Eagly and Chaiken programme, the over-200-studies count, and the entire Earthquake Commission case including the corrupted statistic. Relabelled the Pinker material as stylistic observation and identified the example-principle-application pattern as a WhenNotesFly framework rather than an empirically established signature. Corrected the Geisler description to academic philosophy and replaced the lost-the-capacity claim with the point that experts may find novice readings harder to reconstruct. Reframed Construal Level Theory as adjacent evidence rather than validation of Hayakawa's model, citing Liberman and Trope 1998 and Trope and Liberman 2010 directly. Reframed the financial crisis and Enron as illustrations rather than causal accounts, removed the duplicated Columbia case in favour of a single Intel Pentium example with its non-abstraction factors restored, and removed the unsourced Intel stock-movement figure and industry-standard claim. Labelled the one-floor-above heuristic and the Bessie table as WNF adaptations, separated jargon from familiarity from abstract generality, added that high abstraction is essential for theory, law and strategy so the model reads as fit rather than descend-when-possible, replaced universal wording throughout, rewrote the FAQ which had prescribed a single sequence, and reduced the article by about 47 percent.
Why this is better
The article's central weakness was that a sound and useful linguistic model had been given a research appendix and four case studies that it could not support. One section asserted a Google research finding as fact and then admitted in the same breath that the finding does not exist, which is worse than either omitting it or stating the limitation cleanly. Another carried a visibly corrupted statistic reading 'by 4substantially' on a live page. Corrections remove what cannot be verified, bound the remaining research to what it establishes, restore the non-communication factors in the case studies, and separate abstraction from the adjacent concepts of jargon and familiarity that the article had merged into it.
More by Emir Baycan in Communication
- Correction Why Miscommunication Happens
- Correction Active Listening: What It Is, How It Works, and How to Practise It
- Correction Narrative Transportation: How Stories Influence Beliefs and Behavior
- Correction The Curse of Knowledge: How Experts Struggle to Explain