Spotting fake in the news: an easy task?

Размещено на http://www.allbest.ru/

Размещено на http://www.allbest.ru/

Spotting fake in the news: an easy task?

Olena Gryshchenko, Galyna Tsapro



Fake news is a widespread element of the nowadays news websites. The research focuses on students' abilities to detect fake news stories. 72 % of respondents successfully identified fake texts. The experiment proves that students concentrate on reading texts carefully, check their credibility, facts and pictures that accompany news texts. Students believe that among linguistic features that contribute to creating fake news texts there is repetition of lexemes, illogical structure of narration, exaggeration, confusing numbers. It was also pointed out that photos do not illustrate information given in fake texts.

Keywords: fake, fake news, media discourse, empirical methodology, text analysis.



Грищенко Олена, Цапро Галина. Визначення фейку в новинах: легке завдання?

Фейкові новини є атрибутом сьогоденних медіа джерел, особливо новинних сайтів в Інтернеті. Розпізнання фейку стає все більш нагальною проблемою. У зв'язку цим було проведено дослідження, мета якого - з'ясувати, як студенти-філологи вирізняють фейкові тексти. Респондентам було запропоновано прочитати два тексти, один з яких - фейковий, другий - справжній, визначити тип тексту та відповісти на питання: 1. Що ви робити, щоб розпізнати фейкові новини? 2. Що саме в тексті вказує на його фейкове походження? До експерименту було залучено 90 студентів, яких розподілили на три групи по 30 учасників у кожній. Відповідаючи на перше питання, респонденти вказали, що для виокремлення неправдивої інформації вони уважно читають текст, аналізують прочитане, покладаються на здоровий глузд, перевіряють факти, імена, джерела інформації, аналізують фото/відео, що супроводжує новини. Перші дві групи студентів успішно розпізнали фейковий текст (82 % і 79 % відповідно). Однак у третій групі лише 55 % указали на фейковий пост. 72 % респондентів правильно визначили фейкові новини, виокремлюючи при цьому лінгвістичні засоби, що вказують на неправдивість інформації. До таких маркерів належать: 1) повтор лексичних одиниць та фраз, 2) оціночні прикметники, 3) абсурдність подій, що описуються, 4) показники невизначеності, 5) вірогідність фальшивих імен, 6) наявність граматичних та орфографічних помилок, 7) недоречні фотографії, що ілюструють новинні події. У третій групі всі студенти, визначаючи правдивий текст як фейковий, указали на недоречне вживання метафори в назві статті, яка розповідає про аварію вантажівки з пляшками кока-коли, при цьому описуючи подію як руйнування свят. 47 % з цих студентів, наголошуючи на неправдивості інформації, вказували на різні години, подані у статті, та стан водія від «непостраждалий» до «отримав незначні травми». Необхідно зазначити, що ці респонденти не звернули увагу на той факт, що стаття подає під одним заголовком інформацію про дві аварії, що сталися на цій трасі, що, своєю чергою, вплинуло на сприйняття новин та ускладнило розпізнання правдивого та сфабрикованого новинного тексту.

Ключові слова: фейк, фейкові новини, медійний дискурс, емпірична методологія, аналіз тексту.

fake news disinformation



Introduction

Fake news has become a widespread phenomenon and is considered to be a usual attribute of nowadays' life. The collocation fake news is defined in the MacMillan dictionary as a story that is presented as being a genuine item ofnews but is in fact not true and is intended to deceive people [FN], which clearly indicates that the process of identifying fake in the news is not painless as it requires certain skills to check the given information in order not to be misled by untrue news stories. Though the meaning of the collocation fake news is quite clear from its components with news meaning material reported in a newspaper or news periodical or on a newscast and fake interpreted as false, counterfeit [RSFN], the term fake news used in political/social/legislative fields is seen as ambiguous as it lacks consistent usage and is closely associated with propaganda and disinformation and fake news can be used as a synonym to the phenomena mentioned above [DP 2019, p. 25].

Disinformation is conscious deception contrasting to ignorance, exaggeration, underestimation or accidental deception [ Почепцов 2019b, p. 13]. Disinformation is conventionally distinguished from misinformation by the intent of its disseminator to deceive [Haigh and Haigh 2020, p. 304]. Propaganda sounds a bit different from disinformation as it derives from the Latin verb propagatus meaning `to spread', whereas the dictionary definition partially implies its false nature: information, often inaccurate information, which a political organization publishes or broadcasts in order to influence people [P]. Gill Bennett [2020, p. 246] emphasizes on the fact that there is no clue within the word propaganda itself as to whether the information to be disseminated is factual or inaccurate (whether by design or error), so clearly one function of propaganda may be, in some circumstances, to spread false information. Still, nowadays propaganda is mostly connected with negative perception of imposition of evil images and ideas and mass manipulation, which is rejected by progressively thinking people.

We see disinformation as a synonym to fake news, thus characterizing fake news as a specific integrative type of media text, the main components of which (lie, deceit and manipulation) correspond to and define its main function - manipulative function [Gryshchenko 2017, p. 135].

It seems that nobody is in favour of being deceived or manipulated especially when it is done on purpose. The society's awareness about fake news is demonstrated through a variety of programs and courses available for different age groups. In order to detect how well students whose major is linguistics can identify false information, the research has been carried out aiming at revealing students' abilities to spot fake news as well as to define verbal and non-verbal means which represent fake news.



Theoretical background

Starting to speak about fake news, several questions arise: why there is so much fake news right now and who benefits from distributing fake news. Is all fake news false in the same way?

According to Georgii Pocheptsov [2019а, p. 52], true/fake information is gradable and readers often grade it not trying to check its credibility but relying on sources and distribution. Another point of grading truth is accepting or rejecting information and this process is not often grounded on reasonable arguments but mostly on allegations. Nowadays public discourse is full of opinions, not facts, and «That's just your opinion!» conveys that the given piece of information equals fake news [Dentith 2017, p. 73], it is perceived as fake news, even if `your opinion' is really based on factual information. This means that factual information is not considered in the proper way, the public consumes opinions whether they are true or not, facts are not accepted and welcomed especially if they do not correspond to some groups or individuals' expectations. «That's just your opinion!» or «That's just fake news» [Dentith 2017, pp. 75-76] cover everything that differs from what is conventional for a specific group from the public.

Kelly Born [2017] points out that there are six reasons why we face more fake news today. The first one deals with the democratization of information creation and distribution. The author brings up the fact that access to networks has enabled a great many people to create and share news, or any stories that can be perceived as news, 7 and in this way influence large audience, which can be beneficial on the one hand, while on the other, be quite risky, for example, when journalistic standards of excellence are not followed. The second reason - information socialization - is tightly connected with the first as sharing news from person to person is becoming even easier every day with the development of new webs and devices. Atomization, the third reason, makes it almost impossible to trace the source of information and in this way to distinguish between credible and noncredible ones. Readers are likely to believe not the source of information but those people who share the given piece of news in networks. The fourth component - anonymity in information creation and distribution - is represented by the lack of a brand and by-line, in other words it means that many pieces of news are deprived of sources or names of writers flourishing the ground for news stories created by bots. The fifth element - personalization - is connected with targeted messages that can be adaptive and personalized depending on readers' interests. Sovereignty is the sixth reason and it means that unlike other kinds of media, social networks operate as self-adjusted and controlled; thus, the amount of ads, political ads especially, is almost impossible to determine or govern.

Vasil Sebastian Dancu [2017] believes that fake news is not produced by amateurs writing blogs in social networks but it i s more likely to be a job of well- organized structures supported by those who have power in society.

Norbert Schwarz and Madeline Jalbert [2021, p. 84] indicate that true stories are more difficult to comprehend than false ones because of simplicity of the latter, that is why the researchers offer to present true information in ways that facilitate its fluent processing,... making the font easy to read and the speaker's pronunciation easy to understand, adding photos and repeating key points, which will give truth better chances to survive in today's world of countless information.

Outline of the experiment. There is a lot of information about fake and fake news: how to spot fake, how to detect fake news. There are numerous articles on the Internet, various online courses, different websites, resources and research guides. But people are still deceived by fake information and experience difficulties while defining fake. Often up to ten ways /steps on how to spot fake are offered in online articles and various books/guides [Bounegru 2017; HSFN; Kiely and Robertson 2016]. It seems quite achievable. Among them are the following: developing a critical mindset; considering the source and supporting sources; reading beyond; studying the images; checking the author and web address; checking your biases; analysing the purpose of information; consulting the experts, etc. However, it continues to be a difficult task.

We conducted an experiment in order to understand what exactly governs the way we detect fake information, which linguistic means help us to uncover fake, which verbal and non-verbal means trigger our reactions and boost brain processes. We aimed at the text, language and different means which go first and precede the aforementioned steps. Students who study Linguistics and languages possess necessary competences (language, linguistic, communicative, discursive, pragmatic, sociolinguistic, grammatical, etc.) and are more attentive to language use, so they can spot different linguistic means in a text which influence our perception of it and help to identify whether it is real or fake. Our hypothesis is the following: every text contains certain linguistic means, which on the level of a text allows to spot fake and students studying Linguistics acquire necessary skills working with the text.

The participants of our experiment were 90 university students (aged 19 -22) whose major is Linguistics (English). All the students were divided into 3 groups randomly, 30 students in each. Each group got 2 texts (1 true and 1 fake). Group A had Text 1 (true) [T1] and Text 2 (fake) [T2]; group B -Text 3 (true) [T3] and Text 4 (fake) [T4]; group C got Text 5 (true) [T5] and Text 6 (fake) [T6]. The following text selection criteria were defined: topicality, authenticity, variety of subject matter, the presence of video or photo (optional). The students were offered to read the texts and define which one is fake. The students were asked to answer two questions in the questionnaire. The first one: How do you spot fake and fake news? The second one: What is there in the text which helped you to detect fake? Support your ideas with examples from the text.

In group A 82 % of students defined the fake text correctly, in group B - 79 %, in group C - 55 %. So, in general 72 % could spot the fake text.

Results and Discussion. To the first question How do you spot fake and fake news? the participants wrote that they in order to identify fabricated news stories:

* read them carefully and attentively (52);

* analyse what is being read (43);

* rely on common sense (42);

* use knowledge and expertise (the students underlined that when a person reads a lot, they are knowledgeable and well-informed) (33);

* check facts and named (31);

* check sources (29);

* google information (27);

* study links (27);

* cast doubt on the given information (23);

* compare numbers, dates and proper names with information (21);

* check the correspondence between the photos/video and the information (18).

The participants named the following features that help them to come to the conclusions the text is fake:

* facts or information sound impossible / paradoxical / unreal / unbelievable / incredible / absurd / hyperbolized (74 responses);

* very few facts / absence of facts (65);

* facts are mixed (60);

* details (very few details / not enough details, no details, extra details) (50);

* no proof / no evidence (48);

* not real photos / pictures (47);

* contradictory / controversial facts (40);

* false geographical or proper names (38);

* the usage of doubt words and the words of uncertainty (some, almost, nearly, etc.) (32);

* the story is not logical (illogical, develops not in a logical way) (30);

* the absence of link and date (21);

* the usage of numbers (numbers of people, things; dates) (20);

* funny / ridiculous information (14);

* the amount of time (12);

* lack of quotations (where necessary) (10);

* mistakes in the text (3);

* no sense, senseless (1).

The second part of the experiment involved students' work with the text and their explanations how fake news is created. Three groups of students offered a bit different choice of linguistic means which helped them to identify fake in the news stories.

Group A. The respondents got text 1 «Russian ship carrying 127 passengers gets trapped in ice for days» and text 2 «Arm Bitten Off After Hugging A Great White Shark». In order to analyse and interpret the data we have studied the examples from the texts which were named by the students as those which helped them to detect fake.

In text 1 the participants, who spotted the fake text correctly, paid attention to the following examples (we supported them with the number of references):

* the great white shark he was trying to HUG, even HUGGING the large animals, after being hugged (23);

* they swam with a group of sharks (22);

* one of the sharks SUDDENLY became aggressive (20);

* suddenly became AGGRESSIVE... and suddenly ATTACKED him (19);

* The young man lost a large amount of blood, but was able to get out of the water (18);

* The group shot some incredible footage during the first 30 minutes, as they swam with a group of sharks (16);

* Mr. Waterford decided to dive with some great white sharks in order to improve the marine predator's public image and raise awareness about shark hunting (11);

* 21-year old Darrell Waterford... was participating in a promotional video for Greenpeace (8);

* An environmental activist was almost killed (2);

* 21-year old Darrell Waterford (1);

* «An environmental activist was almost killed Yesterday in the Indian Ocean» (1);

* It attacked Mr. Waterford and tore off his right arm (1);

* suddenly attacked him and bit his arm off (1);

* some 100 nautical miles away from the Australian city of Perth (1);

* incredible footage with sharks (1);

* some great sharks (1).

The study of the mentioned above examples demonstrate that the respondents paid attention to such linguistic means: repetition of words suddenly, hug / hugging / hugged, some; repetitions of actions attacked and tore off his arm, attacked him and bit his arm off; numerals one of the sharks, twenty-one-year-old, one hundred miles, four other members, thirty minutes; markers of uncertainty almost, some, nearly; rapid sequence of events with Past Simple (suddenly attacked him - bit his arm off; suddenly became aggressive - suddenly attacked him); descriptive adjectives to put some emphasis on particular events in the text (incredible, aggressive, dangerous).

A lot of events sound unbelievable and absurd: hugging sharks, swimming with sharks, swim with sharks to improve predator's public image, etc. Time markers are specified within unreal time frames: 30 minutes of swimming with sharks; The young man lost a large amount of blood, but was able to get out of the water and was rapidly taken to the hospital (quickly and impossible). Discursive markers of time and place are used in opposition: far away - near, quick - slow, rapidly - slowly, e. g., 100 nautical miles away from the Australian city of Perth - was able to get out of the water; swimming for 30 minutes - the sharks suddenly attacked.

Moreover, the text contained a mistake: Yesterday - capital letter in the middle of a sentence. And some students paid attention to the hidden face in the photo and the full name, age and occupation of the person shown.

In text 2 the participants, who mistook a true text for the fake one, paid attention to the following examples (we supported them with the number of references):

* has been trapped in ice for days off the coast of Japan (5);

* It's unclear exactly how many days the ship has been trapped (4);

* A Russian ship carrying 127passengers (1);

* it became stuck in ice up to 3 feet thick (1);

* There's no immediate danger to the passengers... for a week (1);

* Now the boat is breaking through it, trying to get out into clear water (1);

* A Russian ship / Russia's Pacific coast / the Russian port (1).

The following examples show that the participants focused on such linguistic means: the usage of numerals 127 passengers, 3 feet thick; markers of uncertainty unclear, for days. Some events seemed unreal and ambiguous: about immediate danger, the boat which could break through the ice but not the ship, the unclear time the ship was trapped. The last reference Russia / Russian is considered to be not a trustworthy source.

Having studied the two texts and the participants' responses, we may come to the conclusions that there are definite linguistic means which are typical for fake texts: repetition of words and actions, rapid sequence of events, descriptive adjectives. Numbers and numerals may have an ambivalent meaning as, on the one hand, they serve as a proof to certain information and events, but on the other hand may be deceptive. Markers of uncertainty are more used in fake texts in order to manipulate and deceive, however some words with the meaning of uncertainty can appear in real texts. The true text does not contain evaluative adjectives, repetition of words or events and does not contain rapid events. Both texts were supported with photos, however the students paid attention to the photo of the fake text.

Group B. The students in Group B received text 3 «Free apple on Delta flight ends up costing woman $500» without a photo and text 4 «White people will be `blended out'» with a photo of Malia Obama.

The participants, who spot the fake text correctly, supported their ideas with such text examples:

* White people are so 1960's. Imagine a world without white people (18);

* they will be blended out by the time I am 30 (11);

* they will be blended out by the time I am 30 (7);

* Imagine a world without white people (4);

* the Facebook page of the America's Last Line of Defense web site posted an image offormer first daughter Malia Obama (2);

* 1 like = 1 arrest (1).

In text 4, a true text, the respondents paid attention to tautology white people, with white as an evaluative and discriminating adjective, imperative mood Imagine.., the usage of numbers I am 30, 1 like = 1 arrest, 1960's, phrasal verbs blend out. Some proper names and information seemed untrustworthy.

In text 4 the students emphasized the following examples:

* A free airline snack wound up costing one woman $500 (5);

* Crystal Tadlock says... (4);

* A free airline snack wound up costing (2);

* She took the piece offruit and put it in her carry-on bag (1).

The participants paid attention to the numbers $500, not a real sum of money for an apple; a full name of a person, which might be fake; phrasal verbs wind up; excessive details in the text that suggest fake actions and events.

According to the information studied above, fake texts contain tautology and repetition of ideas, numbers and numerals, some proper names, which can be fabricated or real names of real people used for manipulative purposes, phrasal verbs and imperative mood. Some detailed and excessive information can seem fake as it has no particular sense in the text. Text 3 was provided with a picture which could influence the respondents' choice.

Group C. In this group the least number of participants identified the fake news text. The students analysed text 5 «The holidays are CRASHING!» and text 6 «Singer Toni Braxton has passed away at age 50 of unknown causes». Both texts were illustrated with photos.

The students (17), who correctly detected the fake text, picked up the following sentences that attracted their attention and made them think that the given information is false (the number of references is provided):

* feeling «unresponsive » at home (14);

* the cause of her death is not immediately known (10);

* amid family feud (9);

* Toni Braxton has passed away at the age of 50 of unknown causes (9);

* Her sister Traci Braxton who was kicked out of her upcoming tour «Sex and Cigarettes » (7);

* after been rushed to hospital (5);

* According to the family's spokesperson Anita Arthur (5);

* singer «showed no sign of illness and never complained of lupus » (4);

* She also appeared on «The Dr. Oz Show» in 2013 to talk about her diagnosis (4).

The linguistic means the respondents paid attention to are the following: the descriptive adjective unresponsive given in quotation marks in the text describing Toni Braxton's state; the time marker immediately used to describe impossibility to give the reason of the death, especially when the disease lupus was mentioned several times in the text, even with explanations of main symptoms; contradictory information about the singer's participation in the show in order to speak about her illness and the singer's sister's quote about absence of any signs of malaise or health disturbances; the specific choice of vocabulary rushed, family feud, was kicked out, which added some extra emphasis to the events described. The students also underlined the use of the Present Perfect in the news title as usually the Present Simple or Past Simple is used. The participants also indicated that in the text there are no dates or hours when the `tragedy' actually took place and considered it as the sign of fake news. Speaking about nonverbal means, in this case, the picture of the singer, the students, addressed to the fact, that there are no pictures of the singer being taken to hospital or in hospital, which in their opinion might justify the piece of news.

All the participants (13 students), who saw the true text as the fake one, wrote about the emotional pounding of the title The holidays are CRASHING! underlying the weird sounding of it, and mentioning that it was the crucial point in defining the text as the fake one. The metaphor in this case was perceived as unnecessary exaggeration of the non-existent tragedy.

Almost all the students (11) doubted that Coca-Cola company carries products of other labels, so they clearly defined the fact that the cargo also included bottles of Schweppes lemonade as the false information.

As the article speaks about two different accidents under one title, the students (7) not being attentive enough, stated that the phrase Nobody was injured in the accident does not correspond to the phrase A man suffered minor injuries in the accident and was taken to Leeds General Infirmary not understanding that two different lorry-drivers were mentioned. The students found this information confusing and as a result the fabricated one. They also pointed out the different hours of the accident in the story 11a.m. and 6.30 a.m. Actually, that was when two accidents happened.

The picture of the lorry lying aside the road with Coca-Cola bottles all over the place seemed too dramatic to students (9) and they characterized the image as photoshopped.

The analysis of linguistic means seen as fake in text 5 and 6 enabled to outline the following results: the irrelevant use of descriptive adjectives, time markers, vocabulary which make an effect of overstating facts and adding unnecessary emotional colouring to the news stories; the use of grammar tense in the title; photos that do not reflect the given stories. Thus, the true news texts include less emotionally charged vocabulary and are more precise and logical in giving information and contain dates to exemplify the occurrence of events.



Conclusions

The research proves that students from Linguistic departments are able to spot fake news using text analysis strategies. The future study is going to offer the same questionnaire and texts to students from other departments in order to compare the strategies they use to detect false news stories with the results obtained.



References

Почепцов, Г. Г. (2019a). Виртуальные войны. Фейки. Харків, 392 с.

Почепцов, Г. Г. (2019b). (Дез)информация. Київ, 248 с.

Bennet, G. (2020). Propaganda and Disinformation: How a Historical Perspective Aids Critical Response Development. [In:] Baines, P., O'Shaughnessy, N. & Snow, N. (eds). The SAGE Handbook of Propaganda, pp. 245-260.

Born, K. (2017). Six Reasons There's More Fake News Today. URL: https://www.marketwatch.com/story/six-features-of-the-age-of-fake-news-2017-10-24 (date of access 15.05.2021).

Bounegru, L., Gray, J., Venturini T. & Mauri M. (2017). Field Guide to «Fake News» and Other Information Disorders. Amsterdam, 216 p.

Dancu, V. S. (2017). Fake News or Fake Concept? [In:] SINTEZA. Review of Culture and Strategic Thinking. 39. Cluj Napoca. URL:

https://www.academia.edu/35894963/FAKE_NEWS_or_FAKE_CONCEPT (date of access 07.04.2021).

Dentith, M. R. X. (2017). The Problem of Fake News. [In:] Public Reason, 8 (1-2), pp. 65-79.

(DP) Disinformation and propaganda - impact on the functioning of the rule of law in the EU and its Member States (2019). URL: https://www.europarl.europa.eu/thinktank/en/document.html? reference=IPOL_STU(2019)608864 (date of access 23.05.2021).

(FN)Fake news. URL: https://www.macmillandictionary.com/dictionary/british/fake-news (date of access 25.05.2021).

Gryshchenko, O. V. (2017). Lie, deception, fake and truth. [У:] Науковий вісник Міжнародного гуманітарного університету. Серія «Філологія», вип. 26, с. 133-135.

Haigh, M. & Haigh, T. (2020). Fighting and Framing Fake News. [In:] Baines, P., O'Shaughnessy, N., & Snow, N., eds. The SAGE Handbook of Propaganda, pp. 303-322.

(HSFN) How to Spot Fake News. URL: https://www.torontopubliclibrary.ca/spotfakenews/ (date of access 25.05.2021).

Kiely, E., & Robertson, L. (2016). How to spot Fake News. URL: https://www.factcheck.org/ 2016/11/how-to-spot-fake-news/ (date of access 28.05.2021).

(P) Propaganda. URL: https://www.collinsdictionary.com/dictionary/english/propaganda (date of access 28.05.2021).

(RSFN) The Real Story of 'Fake News'. URL: https://www.merriam-webster.com/words-at- play/the-real-story-of-fake-news (date of access 15.02.2021).

Schwarz, N. & Jalbert, M. (2021). When (fake) news feels true. Intuitions of truth and the acceptance and correction of misinformation. [In:] Greifeneder, R., Jaffe, M. E., Newman, E. J., & Schwarz, N., eds. The Psychology of Fake News. Accepting, Sharing, and Correcting Misinformation. Routledge, pp. 73-89.

(T1) Russian ship carrying 127 passengers gets trapped in ice for days. URL: http://abcnews.go.com/International/russian-ship-carrying-127-passengers-trapped-icedays/story?id= 52931799&cid=clicksource_4380645_1_hero_headlines_headlines_hed (date of access 15.02.2021).

(T2) Arm Bitten Off After Hugging a White Shark? URL: https://www.snopes.com/greenpeace- activist-get-arm-bitten-off-hugging-white-shark/ (date of access 15.02.2021).

(T3) Free apple on Delta flight ends up costing woman $500. URL: https://abc7news.com/travel/free-apple-on-delta-flight-ends-up-costing-woman-$500/3378567/ (date of access 15.02.2021).

(T4) Malia Obama: White people will be `blended out'. URL:

https://www.politifact.com/factchecks/2018/oct/18/blog-posting/there-no-evidence-malia-obama-said- white-people-sh/ (date of access 15.02.2021).

(T5) The holidays are CRASHING! Coca-Cola truck overturns on the busy M62 spilling bottles of soft drink across the motorway. URL: https://www.dailymail.co.uk/news/article-5662047/The- holidays-CRASHING-Coca-Cola-truck-overturns-M62.html (date of access 15.02.2021).

(T6) Singer Toni Braxton has passed away at age 50 of unknown causes. URL: https://www.snopes.com/fact-check/toni-braxton-death-hoax/ (date of access 15.02.2021).

Размещено на Allbest.ru

...