Unit 1: Five-Filters Approach

Despite the focus on the novelty and substance of research contributions, research articles that are permeated with lexical, grammatical or genre-related errors are more likely to be rejected. Intrusive errors may cause reviewers to misunderstand the intended meaning, severely lowering the chance of acceptance.

One non-intrusive language error is unlikely to result in rejection, but multiple non-intrusive errors may create a negative impression, possibly leading academic gatekeepers to fall victim to the horns-and-halo effect — assuming that the research itself may also contain errors. In short, writing error-free prose reduces the possibility of rejection.

In this course, 22 specific types of language errors are identified, explained, and suggestions for addressing them are given. As there are no native speakers of scientific communication, these errors apply to both native and non-native English speakers.

This is overwhelmingly the most common category of error. Accuracy errors differ greatly in their severity. Errors in grammatical accuracy may or may not affect the meaning of a paper. However, errors in factual accuracy can not only affect the meaning but may destroy it entirely.

Research writing is judged on its merits. Research that is factually inaccurate will be dismissed. A student submitting a graduation thesis that contains falsehoods is unlikely to graduate. Doctoral candidates submitting dissertations with statistical and numerical errors will no doubt need to revise and resubmit.

The accuracy filter covers six subtypes of error, which are explored in detail in Unit 2: factual errors in world knowledge, factual errors within the article, overly bold claims, overgeneralization errors, statistical and numerical errors, and spelling and grammatical errors.

Brevity errors occur when a writer uses more words than necessary. The aim is to convey maximum meaning in the minimum number of words. When there is a choice, writers should select expressions that contain fewer words over those with more.

Genres may be placed on a cline from terse to verbose. Terse genres, such as scientific research articles, are concise but may be considered difficult to read. To quote Faber (2017), "no one wants to read excessively long studies." Words that do not add substance to the meaning should be omitted.

The brevity filter covers three subtypes: using multiple vague words, repeated words, and redundant words. These are explored in Unit 3.

To ensure readers understand the intended meaning, it is necessary to write clearly. Clarity is realized by selecting expressions that are specific and unambiguous. Words that are overly general or vague may confuse readers, since different readers may interpret the same vague word differently.

Words that have two meanings — ambiguous words — are also problematic. The aim of the clarity filter is to avoid vague and ambiguous wording.

The clarity filter covers five subtypes: use of vague expressions, lexical ambiguity, referential ambiguity, syntactic ambiguity, and garden-path sentences. These are explored in Unit 4.

Research writing that is subjective may not be taken seriously in some scientific disciplines. Computer science is one such discipline. This means that computer scientists need to write in such a way that they cannot be accused of being subjective.

This can be achieved by focusing on processes and results, and by removing wording that is personal or emotional. Such writing may not be truly objective, but it will not appear subjective. Objectivity can be considered an idealized aim, but avoiding accusations of subjectivity is the practical goal.

The objectivity filter covers three subtypes: focus on people and feelings rather than things and ideas, emotive wording, and excessive personalization. These are explored in Unit 5.

Formal writing is characterised by the use of standard English. Specifically, spelling conventions are followed, accepted grammatical forms are used, and informal expressions are avoided.

The belief is that formal writing is trustworthy — this is also why newscasters use formal language to read the news. Readers of research articles written using informal language may not trust the content of the research. The convention is not going to change soon, so writers are advised to write in a formal manner.

The formality filter covers five subtypes: apostrophes, abbreviations, slang, informal terms, and rhetorical questions. These are explored in Unit 6.