Eyewitness Testimony and Daya Singh v. State of Haryana - A Critical Reappraisal

This paper is written as a reaction to the judgment delivered by the Supreme Court of India in the case of Daya Singh v. State of Haryana^[1] and the focus of this paper will primarily be on eyewitness testimonies and the inherent flaws in relying heavily upon such testimonies while reaching a conclusion regarding the guilt of an accused. Tracing judicial decisions, it will become obvious that since the very advent of law enforcement, eyewitness testimonies have played a focal and extremely decisive role in the identification, arrest and conviction of suspects. The underlying assumption for the acceptance of eyewitness testimonies as a substantive piece of evidence is that the human mind is skilled and seamless at recording, storing and retrieving information regarding events.^[2]

In the case of Daya Singh v. State of Haryana, the incident in question dated back to 1988 and the controversy in the appeal dealt with the narrow issue of the conviction of the appellant based solely upon the identification of the accused in court by two prosecution witnesses. The appellant, who was the accused in the case, was arrested on 28 May 1988 and the identification parade was to be conducted on 2 June 1988, but on that day the accused refused to take part in the Test Identification Parade (TIP). Eventually, the identification in court by the witnesses took place almost 8 years after the incident had occurred. This paper seeks to address the issue of whether courts as a general practice, and in particular whether the judges in the above mentioned case, should have accepted such evidence and relied upon it as heavily as they did while reaching a conclusion regarding the guilt of the accused.^[3]

The Supreme Court, while passing a sentence for conviction of the accused, stated that, "Courts ought not to increase the difficulties by magnifying the theoretical possibilities. It is their province to deal with matters actual and material to promote order and not surrender it by excessive theorizing or magnifying what in practice is really unimportant."^[4] The fact of the matter however happens to be that it is not a mere theoretical possibility with an extremely low likelihood that the eyewitness testimonies could have been inaccurate. Research on such testimonies and the resultant data collected while examining cases of wrongful convictions elucidates that not relying as heavily as the Court did in this case would in fact classify as extremely prudent behavior. This must not be considered as "magnifying the theoretical possibilities". There is abundant literature available on how eyewitness testimonies are unreliable and this in no stretch should be termed as "excessive theorizing or magnifying what in practice is really unimportant."

While assigning reasons for as to why the bench believed there was nothing on record to cast a doubt on the reliability of the eyewitness testimonies, the Court stated that the prosecution witness had provided specific physiognomy of the accused. This particular physiognomy that the Court talks about was the fact that the witnesses had stated that the accused had "catty eyes" and when asked what they meant by "catty eyes" the prosecution witnesses responded by saying "the eyes like a cat".^[5] This is a prototypical example of using intuitively circular definitions wherein individuals, while attempting to define a particular word, make use of the very same word in the definition itself and this, in my opinion, should not have been held to be adequately descriptive of the features of an individual.

In the words of the Supreme Court, "Where evidence is cogent, consistent and without any motive, it is no use to imagine and magnify theoretical possibilities with regard to the state of mind of the witnesses and with regard to their power of memorizing the identity of the assailants." The Court very evidently failed to consider the immense research on the subject of eyewitness testimonies indicating that the testimonies of even the most honest and best intentioned witnesses is fallible.^[6] Eyewitness testimonies are extremely susceptible to what has been termed as creation of "false memories" and this refers to the tendency of remembering events that never really happened or to remember events as happening quite differently from the way in which they actually happened. The accuracy of these testimonies depends upon the manner of acquisition of information that is to be stored in memory, the duration of retention of this information as well as the retrieval of the information. Along with this, there is an interplay of numerous other variables such as the duration of the event in question, the seriousness of the event, length of the retention interval etc., that affect the accuracy of retrieval of information.^[7]

In a study conducted by Lipton in 1977, subjects were shown a film of an armed robbery and shooting and then asked to recall what they had witnessed. One group of subjects was tested immediately and another group of subjects was tested one week after the event. The results indicated that there was a significantly lower recall value for the group that was tested a week later as compared to the group that had been tested immediately after they were shown the video.^[8] If an interval of one week between an incident and the date of retrieval can lead to significantly inaccurate accounts of eyewitness testimony rendering them unreliable, an identification of an accused by eyewitnesses as in the case of Daya Singh v. State of Haryana, after a retention period of 7-8 years should have definitely raised a serious doubt on the accuracy of the identification made.^[9]

Elizabeth Loftus, an American cognitive psychologist who has specialized in the field of human memory, conducted numerous studies in which college students were presented with a film of a complex event (e.g. a car accident) and immediately afterward were asked a series of questions. Typically, some of the questions were designed to present misleading information, that is, to suggest the existence of an object that did not in fact exist. In one study, subjects who had just watched a film of an automobile accident were asked, "How fast was the white sports car going when it passed the barn while traveling along the country road?" whereas no barn existed. Those subjects were substantially more likely to later "recall" having seen the nonexistent barn than were subjects who had not been asked the misleading question. These experiments, and others that use variations of this procedure, show that memory is an active reconstructive process in which humans interpret and transform information that they encounter.^[10] It involves people picking up information, whether it is true or false, and integrating it into their memory, thereby supplementing and altering their recollection.^[11]

A major eye-opener for the judicial system should have been the findings of a meta-analysis of courtroom trials conducted by Elizabeth Loftus in 1983. The results of this study indicated that testimonies of eyewitnesses were the most persuasive form of evidence and were given significantly higher weightage as compared to fingerprints, polygraph tests and handwriting samples. These results have been backed up by several other studies in different jurisdictions, making the results of the tests more generalized and reliable.^[12]

As per the results from a study conducted by Pezdek, the longer the period for which an eyewitness views a criminal, the higher is the probability of the eyewitness accurately identifying the criminal.^[13] In the case that this paper seeks to address, the eyewitnesses were in a constant scuffle with the criminal and in all probability would have been focusing on defending themselves and preventing the armed terrorist from shooting them as compared to focusing closely on the facial and other features of him in order to make an accurate identification in the future. Additionally, one of the eyewitnesses even ran away and locked himself in a room to call the police which would have resulted in a further reduced period of exposure to the criminal. However, this did not cast any suspicion regarding the reliability and accuracy of the testimony in the mind of the judges.^[14]

Furthermore, witness stress and anxiety induced by a crime can have a negative influence in eyewitness testimony. Environmental factors at a crime scene such as a weapon have the potential of inducing greater levels of stress. This could in turn result in the witness paying greater attention on the weapon and focusing less on aspects such as the facial features of the perpetrator which would have been useful in making an accurate identification during the test identification parade.^[15] These findings seem to be relatable to the case at hand. One of the witnesses went to the extent of mentioning that the criminal was holding a sten gun in his hands but the only description regarding the physical features of the criminal given were that he was well built, had a thick beard and would have been around 25-26 years of age.

Another major inherent flaw of eyewitness testimonies is that witnesses often assume that when a test identification parade is conducted, the actual perpetrator will be a part of the lineup. As a result of this, the eyewitnesses have a tendency to identify from within the lineup the member who most closely resembles the alleged offender as the actual perpetrator. Such fallacies of reasoning have often led to wrongful convictions.^[16]

Furthermore, the eyewitnesses in the case of Daya Singh^[17] were extremely confident about their ability to recall the events with precision and the Supreme Court as well was quite convinced about the veracity of the eyewitness testimonies. With regard to the testimony of one of the eyewitnesses for the prosecution the Court stated that "he (the prosecution witness) has stated that he could identify the appellant after wearing and removing his spectacles". Similarly, in reference to the testimony of the other witness for the prosecution, the Court stated that "her evidence is so natural that it is impossible to believe that she is falsely involving the accused person". The Court found no reason for as to how people with no motive to falsely accuse a person and who were so confident about what they saw, could be inaccurately identifying a person as a culprit.^[18]

While it is no doubt intuitively appealing to believe that an eyewitness who expresses certainty regarding his/her identification of an accused individual (such as in the case of Daya Singh v. State of Haryana) will in all probability also be accurate about it, social psychology experiments and results from numerous studies have shown that the confidence expressed by an eyewitness is indeed a strong determinant of whether people believe that the testimony is accurate. However, what research has not been able to find consistent evidence for is the link between eyewitness confidence and eyewitness identification accuracy. More recent work in psychology continues to show that observers tend to overestimate the diagnostic value of confident identifications, even though confidence is a poor proxy for accuracy outside of carefully controlled, "pristine" identification procedures.^[19] Keeping the above in mind, based upon the confidence exhibited by the prosecution witnesses, the Supreme Court should probably not have been as certain as it appeared to be regarding the veracity of the testimonies in the case of Daya Singh v. State of Haryana.

One aspect of the case that does support the claim of the eyewitness's testimony being accurate is the fact that the event in question relates to an attack by terrorists and could have created a flashbulb memory. This refers to a memory that is encoded during a period of extreme psychological stress and is said to create an extremely accurate and vivid record of the event in one's memory. Justice Shah in his opinion mentioned that "the prosecution witnesses had lost their son, daughter-in-law and the son of their brother-in-law and that it was an extraordinary circumstance for them to be assaulted by terrorists. But it would be difficult to hold that at that time, they had lost their power of perception," and that the "physical features of the accused Daya Singh must have been embedded in the memory of the eyewitness because it was he, who with his two co-assailants committed the gruesome crime".^[20] However, research on this topic has been inconclusive. In fact, more recent studies indicate that flashbulb memories decay at broadly the same rate and involve the use of the same neural mechanisms as ordinary autobiographical memories, resultantly being no more accurate than typical memories. Similarly, flashbulb memories may often be described with greater confidence and more vividly but the reliability of such memories is still in question.^[21]

Wrongful Convictions and the Global Experience

Erroneous eyewitness testimonies have been the single greatest cause for documented wrongful convictions in the criminal justice system of the United States of America.^[22] Data collected from DNA exoneration cases shows that eyewitness misidentification has been involved in roughly two-thirds of such exonerations, making it the leading contributing cause of wrongful convictions there.^[23] Subsequent analyses of these cases reveal that a substantial proportion of exonerees were first placed under suspicion because they resembled a composite sketch or were otherwise suggestively presented to witnesses, which further underlines the dangers of contaminated identification procedures.^[24]

One must also bear in mind that wrongful convictions do not only cause damage to the victims and their immediate families. The trickle-down effect of these convictions can be felt for generations to come. Adding to this is the possibility of additional crimes being committed by the actual perpetrators who will continue to roam around freely in society.^[25]

The fact that hearsay evidence is not received as a relevant piece of evidence could be looked upon as an acknowledgement of the fact that memory traces decay and that what is remembered and conveyed may not be an accurate description of events. It is said that the truth is diluted and diminished with each repetition. Even if not the same, a similar standard, in my opinion, should be adopted with regard to eyewitness testimonies as well.^[26]

Procedural Safeguards in Identification

There are certain measures that must be taken in order to significantly reduce the probability of wrongful convictions while employing eyewitness testimonies. It has been recommended that when a test identification parade is conducted, where there are multiple suspects, only one suspect be included in the lineup at a time. Including numerous suspects increases the probability of a defendant having been selected on the basis of guesswork. The individuals apart from the suspects who are included in the identification parade should be selected in such a manner that they are of the same age range, race and generally fit the description of the perpetrator. The instructions that are tendered to the eyewitnesses should include a statement that the perpetrator may not be present in the lineup and such instructions have been shown to significantly decrease misidentifications.^[27]

Furthermore, there is a possibility of humans subtly and unconsciously communicating information, for instance by gestures and eye movements. Due to this, double-blind techniques should be employed while administering an identification parade, wherein neither the witness nor the officers who are administering the lineup are aware of which person is the suspect in the case at hand.^[28]

Another step that has been widely suggested should be taken is that when a witness selects an individual that he/she believes is the culprit, a record should be made contemporaneously of their degree of confidence. Research has indicated that witnesses' confidence regarding their selections could increase over time and this positive change in confidence levels could later be a determining factor while considering the credibility of an eyewitness's testimony.^[29]

Use of a sequential lineup rather than a traditional lineup has proven to reduce errors of identification. In a traditional lineup witnesses have a tendency of comparing people and selecting the individual who resembles the culprit, even in cases wherein the real culprit is absent from the lineup. In a sequential lineup, people or pictures are revealed to the eyewitnesses one at a time. As a result of this, the witnesses end up focusing exclusively on the person presented rather than resorting to a method of comparison.^[30]

Recent Indian Legal Developments (2018-2025)

Since the decision in Daya Singh, there have been important statutory and judicial developments in India that directly affect how courts should approach eyewitness identification. Section 54A of the Code of Criminal Procedure, 1973 was inserted to empower courts to direct that an arrested person be produced for identification by witnesses, thereby giving an explicit statutory footing to Test Identification Parades and their proper conduct.^[31] More recently, courts and commentators have emphasised that the primary functions of a TIP are (a) to test the memory and credibility of a witness who claims to have seen the accused and (b) to reduce the risk of suggestive or contaminated identifications, rather than to mechanically justify a dock identification.^[32]

A significant change has also occurred in the area of witness protection. In Mahender Chawla v. Union of India the Supreme Court approved the Witness Protection Scheme, 2018 and directed that it operate as "law" under Articles 141 and 142 of the Constitution until a comprehensive statute is enacted.^[33] The Scheme provides for graded protection measures, in-camera proceedings and Vulnerable Witness Deposition Complexes in district courts, and is aimed at ensuring that witnesses can depose without intimidation or fear of retaliation. Although primarily focused on safety, these measures can also indirectly improve the quality of eyewitness testimony by reducing the stressors and pressures that further distort memory.^[34]

At the level of case law, recent Supreme Court decisions have displayed a growing scepticism towards bare dock identifications and flawed TIPs. In a 2025 decision, the Court held that dock identification without a prior test identification parade is generally an unreliable basis for conviction when the witness had no prior familiarity with the accused, insisting that some reliable prior identification procedure is ordinarily necessary.^[35] In another judgment delivered the same year, the Court ruled that where a witness has had an opportunity to see the accused before the TIP, the proceedings lose much of their reliability and cannot be treated as strong corroborative material.^[36] At the same time, the Court has reiterated that failure of a TIP is not automatically fatal to the prosecution, and that courts must examine the totality of circumstances surrounding the crime, the witness's opportunity to observe, and the consistency of their account before either relying on or discarding identification evidence.^[37]

Taken together, these developments mark a discernible shift away from the relatively uncritical reliance on eyewitness identification that characterised judgments such as Daya Singh, and towards an approach that is more sensitive to the empirical research on the fallibility of human memory and the dangers of suggestive lineup practices.

Contemporary Empirical Research and Technology

In the decades since Daya Singh, cognitive scientists have substantially deepened our understanding of how and why eyewitness memory fails. Recent reviews emphasise that memory is best understood as a reconstructive, not reproductive, process, and that contamination can occur at every stage�encoding, storage and retrieval�especially where repeated questioning, media coverage or suggestive procedures are involved.^[38] Modern experimental work has continued to show that exposure to misinformation, even when people are warned about it, can significantly distort later recollection of critical details of an event.

Large-scale analyses of exoneration databases have strengthened, rather than weakened, the claim that eyewitness error is central to wrongful convictions. Organisations such as the Innocence Project continue to report that eyewitness misidentification features in a clear majority of DNA exoneration cases, typically in the range of 60-70 per cent of such cases.^[39] These figures now span several hundred exonerations and cut across offences and jurisdictions, making it increasingly difficult to dismiss wrongful identifications as isolated aberrations.

On the issue of confidence, newer studies have confirmed that decision-makers (judges, jurors, police and laypersons) still rely heavily on the apparent confidence of an eyewitness when evaluating credibility, notwithstanding the weak and inconsistent relationship between confidence and accuracy in most field-like conditions.^[40] These findings reinforce the criticism that the Daya Singh bench placed undue emphasis on the witnesses' asserted certainty and lack of apparent motive to falsely implicate the accused, while underestimating the cognitive limits and biases that affect even well-meaning observers.

Recent research has also begun to explore whether carefully designed technological tools can assist in reducing wrongful identifications, rather than replacing human judgment altogether. Experimental work published in 2025 suggests that AI-based assistance can, in some settings, improve people's ability to distinguish more reliable from less reliable identifications or faces, potentially offering one supplementary safeguard if used in a transparent and rigorously validated way.^[41] At the same time, parallel work on deepfakes and AI-generated faces warns that technology can also create new risks of deception and misidentification, underscoring the need for strong procedural controls when digital images or AI tools play any role in investigations.

Conclusion

I am not saying that there was a miscarriage of justice in the case of Daya Singh v. State of Haryana. However, at the same point in time as a law student with a background of psychology, it is hard not to have grievances with the manner in which the Supreme Court reached the conclusion regarding the guilt of the accused. The judgment highlights numerous fallacies of reasoning while it fails to take into account and give adequate weightage to the vast amount of literature that indicates eyewitness testimonies are far from being infallible and on numerous occasions can be contaminated and made to produce incorrect reconstruction of the crime, leading to the conviction of innocent individuals while the guilty party goes free.

A judge must at all times keep in mind the highly malleable nature of human memory and the flaws in the lineup practices that are employed by law enforcement agencies. Likewise, it is essential for courts and law enforcement officials to be made adequately aware of recent developments on the issue of eyewitness testimonies in the field of social psychology and cognitive science in order to reduce the number of wrongful convictions. The subsequent evolution of Indian law�from the insertion of Section 54A CrPC and the Witness Protection Scheme, 2018 to the Supreme Court's recent scepticism of unsupported dock identifications�shows that the system is slowly, but surely, moving in this direction. The task now is to ensure that trial courts consistently internalise these lessons so that cases like Daya Singh are approached with the degree of caution that modern science and experience demand.

References:

1. Daya Singh v. State of Haryana, (2001) AIR SC 1188.
2. Manveen Singh, "In Eyes, We Trust: The Changing Landscape of Eyewitness Testimony" (2017) 37 Northern Illinois University Law Review 444
3. Daya Singh v. State of Haryana, (2001) AIR SC 1188.
4. Ibid.
5. Ibid.
6. Elizabeth F. Loftus, "Eyewitness Testimony: Psychological Research and Legal Thought" (1981) 3 Crime and Justice 109
7. Henry L. Roediger, "Misinformation Effects in Recall: Creating False Memories through Repeated Retrieval" (1996) 35 Journal of Memory and Language 317-318
8. Vladimir J. Konečni and Ebbe B. Ebbesen, "Courtroom testimony by psychologists on eyewitness identification issues: Critical notes and reflections." (1986) 10 Law and Human Behavior 11118-121
9. Daya Singh v. State of Haryana, (2001) AIR SC 1188.
10. Emory Smith, "The Fallibility of Eyewitness Testimony" (2003) 1 The American Journal of Police Science 492-495
11. Elizabeth F. Loftus, "Eyewitness Testimony: Psychological Research and Legal Thought" (1981) 3 Crime and Justice 109
12. Margaret J. Lane, "Eyewitness Identification: Should Psychologists Be Permitted to Address the Jury?" (1984) 75 The Journal of Criminal Law and Criminology 1325-1328
13. Mark H Ashcraft and Gabriel A Radvansky, Cognition (Pearson 2010)
14. Daya Singh v. State of Haryana, (2001) AIR SC 1188
15. Daya Singh v. State of Haryana, (2001) AIR SC 1188
16. Gary L. Wells, "Effects of expert psychological advice on human performance in judging the validity of eyewitness testimony." (1980) 4 Law and Human Behavior 284-285
17. Daya Singh v. State of Haryana, (2001) AIR SC 1188.
18. Ibid.
19. Gary L. Wells and John W. Turtle, "Eyewitness testimony research: Current knowledge and emergent controversies." (1987) 19 Canadian Journal of Behavioural Science/Revue canadienne des sciences du comportement 384-388
20. Ibid.
21. Mark H Ashcraft and Gabriel A Radvansky, Cognition (Pearson 2010)
22. Richard Wise and others, "A Tripartite Solution to Eyewitness Error" (2018) 97 The Journal of Criminal Law and Criminology 813
23. "Eyewitness Misidentification - Innocence Project", 2018
24. Manveen Singh, "In Eyes, We Trust: The Changing Landscape of Eyewitness Testimony" (2017) 37 Northern Illinois University Law Review 454
25. Vepa P Sarathi and Abhinandan Malik, Vepa P. Sarathi's law of evidence (7th edn, Eastern Book Company 2017)
26. "The Trouble with Eyewitness Identification Testimony in Criminal Cases", 2018
27. Ibid.
28. Ibid.
29. Ibid.