10

Both German and Norwegian has terms for the case when one error is the result of another error ("Folgefehler" and "følgefeil") respectively. As an example of what I mean, you would have a "Folgefehler" if you answered a math question wrong, even if your algorithm was correct, because you relied on answers from a previous calculation that was wrong.

A recurring theme is that speakers of both languages have issues in finding a good equivalent term in English. Is there one?

One suggested translation is "subsequent faults" (alternatively using "errors/defects"), but after googling it does not really seem to be used all that much and it does not seem to fit very well. Another version on a translator site suggests "consequential error", which I think fits better from how I think of its intended use, but is also not a common term. Of course, in this day and age I also asked ChatGPT for a suggestion, but its suggestions were pure hallucinations :D suggestions from gpt-3.5

So far I have not been able to do better than consequential error, but there must be some kind of semi-known term for this?

9
  • 7
    What is hallucinogenic about the answer from ChatGPT? At least two answers include “Cascading Errors”. Recommend you edit your question to sound less flippant.
    – user177197
    Commented Sep 22, 2023 at 14:26
  • 2
    Compared to cascading errors, I think that a "cascade of errors" insists less on the causal relationship between the events.
    – Graffito
    Commented Sep 22, 2023 at 15:39
  • 1
    Forget Google for this type of thing.
    – Lambie
    Commented Sep 22, 2023 at 22:54
  • 10
    To me, consequential error primarily suggests an error that has important consequences, rather than one that is a consequence. Commented Sep 22, 2023 at 23:37
  • 3
    "Error propagation" is also a real term. Commented Sep 23, 2023 at 7:49

7 Answers 7

16

In your specific context, this is known as an "error carried forward" (which is apparently so common in grading papers that it also can refer to a grading scheme in which those wrong answers can still count for some points in subsequent questions).

For example:

(C grade) You can have an error carried forward for your value of m (or r) from (a)(ii) or (a)(i). If you use the symbol e instead of Q (or q) you can get a maximum of only 2 marks. — AQA A-level Year 2 Physics Student Guide: Sections 9 and 12

See also Wikipedia:

Error Carried Forward (ECF) is an informal principle for grading employed within computational fields of study such as mathematics, physics, engineering and computer science. In questions with multiple parts, it is common that the answer to the current part builds on an answer to the previous part. As such, if the answer to any part is incorrect, all subsequent parts will be incorrect, even if the approach for said subsequent parts was correct. The purpose of Error Carried Forward is to protect students who run into this issue from being penalized not only for the initial error, but for all subsequent errors that are only incorrect in answer, not approach.

1
  • 2
    While there are other good answers for the title question, this looks to me to be the best answer for the particular context that was explicated in the body of the question. Commented Sep 24, 2023 at 6:41
13

I call them...

cascade errors
a series of escalating errors

Over 50 years ago I was a novice programmer writing code on punched cards that the computer operators fed into the mainframe computer overnight. I missed out the full stop after DATA DIVISION in a COBOL program, which wasted an entire 2500-sheet box of listing paper on compiler cascade errors before an operator about to load another box of paper into the line printer realized that they might as well just kill the program!

(The computer operators encouraged me to make more such errors, because they ran a "Christmas party slush fund" selling used computer listing paper for recycling, but my boss wasn't at all happy! :)

8
  • 1
    And now that waste of physical resources from decades ago is somewhat redeemed by the value of the story you've shared - an easily understood example of tech going into a haywire domino effect. Everything is information. Nothing is lost, only transformed. And the universe re-balances itself, even if over time scales our feeble human minds will never comprehend. Just contemplating it sends me into cascade errors.
    – Mentalist
    Commented Sep 23, 2023 at 21:34
  • I think compiler-writers learned their lesson, to not output tens of thousands of errors. And, I miss those old line printers. :) Commented Sep 24, 2023 at 7:51
  • @NickGammon you'd be surprised. I did something similar as as FORTRAN tyro, and there's a class of programmer that prefers to blame the undergraduate rather than fix the compiler and (locally-written) job control system. Commented Sep 24, 2023 at 11:27
  • @MarkMorganLloyd When I was young (many years ago) we used a student-oriented Fortran compiler called Minitran. A job was limited to, as I recall, 10 seconds run time, or 50 pages of output. Due to a programming bug I once used all 50 pages. (I did a new page after reaching line 50, and forgot to reset the line number to 1, so each subsequent line went on a new page). Commented Sep 24, 2023 at 21:19
  • 1
    @NickGammon, these days, compilers don't output tens of thousands of errors, they output errors tens of thousands of pages long. (Seriously, just try compiling something using GCC when you use the equality operator on two things that can't be compared for equality. It produces something like twenty pages of candidate type conversions with explanations on why each one can't be used.)
    – Mark
    Commented Sep 25, 2023 at 6:50
10

Not specific to errors, and not very formal, but you could call that a knock-on effect:

A secondary, often unintended effect; a repercussion, chain reaction.

(Wiktionary notes that as a British usage, though it hadn't occurred to me that it might not be equally common elsewhere.)

2
  • Google gives a fair number of relevant hits for ‘knock-on errors’, which sounds perfectly logical and idiomatic to me. Commented Sep 24, 2023 at 22:24
  • I sometimes hear or see "knock-on error" which is strictly speaking a malapropism / contamination, but one that I can get behind since the practice of computer coding clearly needs the idiom!
    – Deipatrous
    Commented Sep 25, 2023 at 12:32
8

Cascading errors is a good phrase for what you are describing. There are two more which have slightly different uses.

The Domino Effect has been extensively used to describe political situations. During the Vietnam War, pundits described the inevitability of all countries in Southeast Asia “going Communist” if North Vietnam won the war. It was compared to a row of dominoes standing on end. Vietnam was the first domino and if tipped over the rest would follow suit.

The Snowball Effect a small snowball rolling down a snow covered hill gathering more snow to become bigger. This implies a continually increasing effect, not a series of discrete steps. It also implies inevitability, once the snowball gets big enough. However, it can be used to predict a negative or positive outcome, depending on the situation being described.

3

In Australian examiners' reports, the name consequential error, which you suggested, is used for an error in an answer when the answer has been worked out correctly from previous answers, but is the wrong answer because of those previous errors.

4
  • Imho those Australian examiners should devote more effort to revising their tests to reduce or avoid situations where the exam candidate might be wasting his time trying to answer a question using incorrect information from a previous question. Apart from anything else, some students might be intimidated because they know they couldn't correctly answer a preceding question upon whose answer the current question depends, where others might be blissfully unaware that they're about to perpetuate a "consequential error". Arguably they're testing confidence rather than competence. Commented Sep 23, 2023 at 16:09
  • @FumbleFingers This issue occurs in the UK too, maybe in most areas of the globe
    – FShrike
    Commented Sep 23, 2023 at 22:32
  • Yes, it's lazy exam design. It would be better if the subsequent questions relied upon a slightly different version of the first question. Say, "if the answer to the previous question was 42 (which it is not) then compute x y z". Commented Sep 24, 2023 at 7:55
  • Marks are awarded if the consequential answer is correct, ie if it follows from the previous incorrect answer. Of course in most Maths exams the majority of the marks are awarded for the working anyway.
    – Peter
    Commented Sep 24, 2023 at 9:22
3

I would call them "compounded errors." "Cascading errors" also sounds good. Note the use of the word "cascade" in "Cascading Style Sheets."

2

Garbage in, garbage out or GIGO is a term used to describe situations in which poor quality input results in poor quality output. This would be appropriate for a situation in which you use the correct method but with input that is wrong. It connotes the sense of carrying errors forward, indicating that subsequent steps can't correct problems that have already occurred. It's often used in the field of data science, to describe the notion that you can't draw meaningful conclusions from data that is flawed to begin with.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .