FPR · First Precedent Research
2026 FPR EDC 17 · https://firstprecedent.org/doctrines/falsus-in-uno
DC-0118Doctrines · Evidence · Editorial draft · 2026 FPR EDC 17
FPR Editors31 October 2026 · 3,814 words
Falsus in Uno, Falsus in Omnibus
The maxim falsus in uno, falsus in omnibus is not a rule of law or practice in India: a court must separate truth from falsehood in a witness's evidence and may convict on the residue, but where grain and chaff are so inextricably mixed that separation would require reconstructing a new case, or where the substratum of the prosecution case is disbelieved, the court cannot build a story of its own.
Abstract
Indian courts rarely see a witness whose evidence is wholly free of exaggeration. Rejecting everything a witness says because one part is false would end many genuine prosecutions; accepting anything that seems plausible would endanger the innocent. This entry examines the maxim's rejection in Ugar Ahir v. State of Bihar, Sohrab v. State of Madhya Pradesh and Gangadhar Behera v. State of Orissa; the limit on sifting in Balaka Singh v. State of Punjab; the classification of witnesses in Vadivelu Thevar v. State of Madras; the concept of the sterling witness in Rai Sandeep v. State (NCT of Delhi); and how these rules fit together in appellate review.
I. The maxim
What it says
1Falsus in uno, falsus in omnibus means false in one thing, false in everything. As a proposition about witnesses, it says that someone shown to have lied about one matter should not be believed on anything else. As an inference, it has some force: a person who has deliberately lied once has shown a readiness to lie. As a rule of law, requiring a court to reject the whole of a witness’s testimony, it is a different matter altogether.
The Indian answer in one sentence
2The Supreme Court’s answer is found in Ugar Ahir v. State of Bihar, in a passage cited ever since:
The maxim falsus in uno, falsus in omnibus (false in one thing, false in every thing) is neither a sound rule of law nor a rule of practice. Hardly one comes across a witness whose evidence does not contain a grain of untruth or at any rate exaggerations, embroideries or embellishments. It is, therefore, the duty of the court to scrutinise the evidence carefully and, in terms of the felicitous metaphor, separate the grain from the chaff.Ugar Ahir v. State of Bihar, AIR 1965 SC 277 (K. Subba Rao J.).
3The passage does three things. It rejects the maxim as law and practice. It offers an empirical reason: exaggeration is common in witnesses. And it imposes a duty on the court. That duty is the real content of the Indian rule, and the rest of the case law is concerned with its proper limits.
II. The rule of caution
Weight, not admissibility
4Gangadhar Behera v. State of Orissa is a strong later statement. The accused in that case argued that, because some witnesses’ evidence had been rejected for certain co-accused, the whole prosecution case should be thrown out. Pasayat J. called the plea “clearly untenable”:
The maxim “falsus in uno falsus in omnibus” has no application in India and the witnesses cannot be branded as liar. The maxim “falsus in uno falsus in omnibus” has not received general acceptance nor has this maxim come to occupy the status of rule of law. It is merely a rule of caution. All that it amounts to, is that in such cases testimony may be disregarded, and not that it must be disregarded. The doctrine merely involves the question of weight of evidenceGangadhar Behera v. State of Orissa, (2002) 8 SCC 381 (Arijit Pasayat J.).
5The language about rule of caution is commonly traced back to Nisar Ali v. State of Uttar Pradesh, AIR 1957 SC 366. The important words are “may” and “must”. A court that finds a witness untruthful on one point is free to disbelieve the rest. It is not obliged to do so. The question goes to weight.
Acquittal of co-accused
6The most common application is where some accused are acquitted on the same evidence that convicts others. Gangadhar Behera says that “Even if major portion of evidence is found to be deficient, in case residue is sufficient to prove guilt of an accused, notwithstanding acquittal of number of other co-accused persons, his conviction can be maintained.” Over-implication of relatives or rivals alongside real assailants is familiar to Indian criminal courts. The rule that a court must reject the entire testimony because some named persons were wrongly accused would hand complete acquittal to the actual offenders.
7The reasoning has an evident risk. A witness who names an innocent person has been shown to be willing to take away a person’s liberty on a false basis. Such a finding must properly make the court more careful about the rest. The rule of caution serves that purpose, and the cases do not treat the acquittal as irrelevant. The question is what evidence remains, and whether the court can rely on it.
III. The limit on sifting
Sohrab: the substratum
8The duty to sift does not mean a court can pick any pieces it likes. Sohrab v. State of Madhya Pradesh states the limit:
Where, however, the substratum of the prosecution case, or a material part of the evidence, could not be belived, it would not be permissible for the Court to reconstruct a story of its own out of the rest.Sohrab v. State of Madhya Pradesh, AIR 1972 SC 2020, (1973) 1 SCR 472 (P. Jaganmohan Reddy J.), reported headnote.
9The same judgment gives a practical reason for the general tolerance of errors. In most cases, witnesses asked about details “venture to give some answer, not necessarily true or relevant for fear that their evidence may not be accepted in respect of the main incident”. That sort of untruth is common and does not prevent reliance on the rest. A false core is another thing.
Balaka Singh: grain and chaff inseparable
10Balaka Singh v. State of Punjab gives the leading formulation of this limit:
Where the grain cannot be separated from the chaff because the grain and the chaff are so inextricably mixed up that in the process of separation the Court would have to reconstruct an absolutely new case for the prosecution by divorcing the essential details presented by the prosecution completely from the context and the background against which they are made, then this principle will not apply.Balaka Singh v. State of Punjab, (1975) 4 SCC 511, AIR 1975 SC 1962 (Syed Murtaza Fazalali J.).
11The Court went on to find that on the facts of the case the evidence of the prosecution witnesses, being partisan and interested, could not be split in that way. The rule therefore has two sides. Separation is required where it is possible. It is forbidden where it would require a new case.
Ugar Ahir on its facts
12Ugar Ahir itself, which is usually cited for the rejection of the maxim, ended in acquittal. The same paragraph that describes the duty of separating grain from chaff also says that the court “cannot obviously” do so in every case. On its facts, the Court found that the evidence on the main issue, including the witnesses’ stated character as disinterested spectators, had been disbelieved. In a phrase worth remembering, the Court said: “To reverse that metaphor, the courts removed the grain and accepted the chaff and convicted the appellants.”Ugar Ahir, AIR 1965 SC 277. The appeal was allowed and the appellants were directed to be set free.
13So the decision that rejects the maxim is also a warning. The practice it demands is a discipline, one that an appellate court will check. A trial court that merely states the metaphor and then proceeds to accept what it wishes has not done what the decision demands.
IV. Witness quality
Vadivelu Thevar
14A related framework comes from Vadivelu Thevar v. State of Madras. The principal issue was whether a conviction may rest on a single witness, and the Court held that it may, since quality counts and not quantity. In the course of that discussion, Sinha J. classified oral testimony:
Generally speaking, oral testimony in this context may be classified into three categories, namely: (1) Wholly reliable. (2) Wholly unreliable. (3) Neither wholly reliable nor wholly unreliable.Vadivelu Thevar v. State of Madras, AIR 1957 SC 614, 1957 SCR 981 (B.P. Sinha J.).
15In the first category, the Court may convict or acquit on a single witness who is above reproach or suspicion of interestedness, incompetence or subornation. In the second category, the court equally has no difficulty. “It is in the third category of cases, that the court has to be circumspect and has to look for corroboration in material particulars by reliable testimony, direct or circumstantial.”
How the framework fits the maxim
16The classification and the rule on sifting are compatible. Most witnesses with some falsehood in their evidence fall into the third category. For such a witness the maxim does not end the matter, and the court does not simply accept the witness either. It looks for corroboration on material particulars. A witness who has been shown to have lied on a material point has moved away from the first category, and the court’s trust must be earned by other evidence. This is how Indian law actually gives effect to the common-sense insight behind the maxim.
The sterling witness
17Rai Sandeep v. State (NCT of Delhi) describes what a witness in the highest category looks like:
In our considered opinion, the ‘sterling witness’ should be of a very high quality and caliber whose version should, therefore, be unassailable. The Court considering the version of such witness should be in a position to accept it for its face value without any hesitation.Rai Sandeep v. State (NCT of Delhi), (2012) 8 SCC 21 (Bench: Swatanter Kumar and Fakkir Mohamed Ibrahim Kalifulla JJ.).
18To test such a witness, the Court said, the status of the witness is immaterial, and what is relevant is the truthfulness of the statement and its consistency “right from the starting point”. The case is helpful because it sets a high bar for conviction on one witness’s word alone, in a case where the Court examined how the prosecutrix’s evidence departed from the FIR. Its standard also makes clear how rarely the first category of Vadivelu Thevar will be available where a witness has been caught in untruth.
V. Minor discrepancies
Not every error is falsehood
19Sifting must also be distinguished from dealing with minor discrepancies. A witness may be simply mistaken. Perceptions and memories differ, and details fade. Bharwada Bhoginbhai Hirjibhai v. State of Gujarat addresses the reappreciation of evidence in the context of minor discrepancies.Bharwada Bhoginbhai Hirjibhai v. State of Gujarat, (1983) 3 SCC 217, AIR 1983 SC 753 (M.P. Thakkar J.). The maxim concerns falsity. An innocent inconsistency does not even raise the question of whether to apply it. A court should identify first whether what it has found is a real falsehood or merely a lapse. Only the former raises the concerns considered in this entry.
Embellishment and the rule against reconstruction
20The boundary between embellishment and fabrication has to be judged case by case. A witness who adds to a true story some small details that did not happen has embellished it. A witness who invents the identity of an assailant or the manner of the attack has fabricated it. The maxim does not apply in the first case. In the second, the fabrication may go to the root of the prosecution case, and the limit in Sohrab and Balaka Singh then comes into play. There is no formula for this. The decisions look to whether what is left can stand as a coherent account of the same occurrence that the prosecution itself put forward.
VI. On appeal and in practice
Duty to give reasons
21A trial judge who accepts part of a witness’s testimony and rejects another part should explain why. The appellate court is bound to check whether the separation was genuinely possible. In Ugar Ahir, the appeal succeeded because the Supreme Court could see that the courts below had not done the exercise in a sound way. The more interested the witnesses and the larger the rejected portion, the stronger must be the reasoning for what is retained.
Advice to counsel
22For the defence, it is seldom enough to show that a witness has lied about something. The maxim will not by itself carry an acquittal. The real target should be to show that what is false is at the core of the case, or that the true and the false cannot be separated without a new story. Cross-examination should be planned with that aim.
23For the prosecution, a clear presentation of what independent evidence supports the retained parts of the testimony is the best answer. Where over-implication appears from the record, it should be faced directly. Pretending that every witness is a sterling witness invites the appellate court to look more closely.
The treatment of the witness who lies
24Rejecting the maxim as a rule of evidence does not excuse a witness’s lying. The separate provisions on false evidence remain. A court which finds that a witness has deliberately lied may consider proceedings for perjury under the applicable procedural law. The law’s tolerance of a mixed witness relates to whether the evidence is of use, not to whether the conduct is acceptable.
VII. Further applications
What the court in Ugar Ahir actually did
25The phrase “separate the grain from the chaff” is often cited without the sentence that follows it. Ugar Ahir went on: “But. it cannot obviously disbelieve the substratum of the prosecution case or the material parts of the evidence and reconstruct a story of its own out of the rest. That is what the courts have done in this case.”Ugar Ahir, AIR 1965 SC 277. The Court listed what had been disbelieved: the pursuit, the assault on the deceased with lathis, the accused going on a bicycle, the deceased wresting a bhala from one of the appellants, the case that the accused attacked the witnesses, and the assertion that the witnesses were disinterested spectators. Then it asked, “If all this was disbelieved, what else remained?”
26That question can serve as a working test. The court should write out what remains of the case after the rejected parts are taken away, and see whether it is the same occurrence that the prosecution described. If what remains is an account the prosecution never gave, the separation has gone too far.
Over-implication in Balaka Singh
27Balaka Singh shows how acquittal of some accused can damage the case against others. The witnesses had implicated nine persons, and the High Court had given reasons for acquitting four of them. The Supreme Court allowed the appeal of the five who were convicted. It considered that the partisan and interested witnesses could implicate the appellants and the four accused equally, and it found it not possible to reject the case as to the four and accept it as to the five remaining appellants. Its reasoning is memorable: “If all the witnesses could in one breath implicate the four accused who appear to be innocent, then one cannot vouchsafe for the fact that even the acts attributed to” the appellants may not have been conveniently made to suit the prosecution.Balaka Singh, (1975) 4 SCC 511. Animus between the parties was an important part of that reasoning.
28Gangadhar Behera reaches the opposite result on its facts, and both are good law. It says that “Where chaff can be separated from grain, it would be open to the Court to convict an accused notwithstanding the fact that evidence has been found to be deficient to prove guilt of other accused persons. Falsity of particular material witness or material particular would not ruin it from the beginning to end.”Gangadhar Behera, (2002) 8 SCC 381. The difference between the cases lies in whether there is something to distinguish those convicted from those acquitted: independent evidence, a specific role, a weapon, an injury, or a recovery. Where the same words are used about everyone and nothing else separates them, the court’s difficulty is obvious. A court convicting some and acquitting others should say what makes the difference.
Why witnesses add detail
29Sohrab gives a practical observation about Indian witnesses. Witnesses asked about details, it said, “venture to give some answer, not necessarily true or relevant for fear that their evidence may not be accepted in respect of the main incident which the have witnessed but that is not to say that their evidence as to the salient features of the case after cautious scrutiny cannot be considered”.Sohrab, AIR 1972 SC 2020. The passage suggests a sensible focus for appraisal. A judge should distinguish the salient features of a case from the details gathered around it. Falsity on the latter may be explained by nervousness or a desire to be believed. Falsity on the former is a much more serious matter.
30This does not excuse a court from noticing how a story has grown. A comparison of the first information report, the statement to the police and the evidence at trial will often show where details were added. An omission of a major fact from the earliest account can amount to a contradiction. Bharwada Bhoginbhai Hirjibhai is usually cited for the opposite warning, that minor discrepancies that do not go to the root should not lead a court to discard evidence, and it is a useful corrective to an overly technical approach.
Consistency and the single witness
31The maxim’s practical importance is greatest where a conviction may rest on one person’s word. Rai Sandeep says of the sterling witness that “What would be more relevant would be the consistency of the statement right from the starting point till the end”, that the witness “should be in a position to withstand the cross-examination of any length and howsoever strenuous it may be”, and that the version “should have co-relation with each and everyone of other supporting material such as the recoveries made, the weapons used, the manner of offence committed, the scientific evidence and the expert opinion”.Rai Sandeep, (2012) 8 SCC 21. A witness who has been found untruthful on a material point can hardly meet that standard. The conviction will then need other support, as Vadivelu Thevar contemplates for the third category.
The contrast with other systems
32Jury directions in some common law systems have traditionally told jurors that they may disregard the testimony of a witness who has wilfully lied about a material fact, in line with the maxim as Wigmore discusses it. Even there the rule has generally been permissive. The Indian position, as Gangadhar Behera frames it, is similar in substance. Testimony may be disregarded, not must. The main difference is that an Indian judge decides both fact and law, and must give reasons, so the application of the rule can be examined on appeal.
The statutory setting
33The Indian Evidence Act, 1872, and now the Bharatiya Sakshya Adhiniyam, 2023, contain no rule requiring a court to reject a witness as a whole. A fact is proved when the court believes it to exist or considers its existence so probable that a prudent person ought to act on that supposition. That definition leaves the weight of evidence to the court. The maxim, if it were a rule of law, would cut across this general approach. The maxim survives only as a caution that a prudent person would bear in mind.
VIII. Weight and discrepancies in the leading cases
Gangadhar Behera in full
34The statement in Gangadhar Behera deserves to be read as a whole, since the sentences leading to the conclusion on the maxim are often left out:
Even if major portion of evidence is found to be deficient, in case residue is sufficient to prove guilt of an accused, notwithstanding acquittal of number of other co-accused persons, his conviction can be maintained. It is the duty of Court to separate grain from chaff. Where chaff can be separated from grain, it would be open to the Court to convict an accused notwithstanding the fact that evidence has been found to be deficient to prove guilt of other accused persons. Falsity of particular material witness or material particular would not ruin it from the beginning to end.Gangadhar Behera v. State of Orissa, (2002) 8 SCC 381, AIR 2002 SC 3633.
35The practical force of this passage is that a mass acquittal of co-accused does not itself affect those who are left. A court may hold that a witness is unreliable as to one person but reliable as to another, because the question is what the rest of the evidence supports.
Bharwada Bhoginbhai
36The basic reasons why discrepancies are to be expected were set out by M.P. Thakkar J. in Bharwada Bhoginbhai Hirjibhai v. State of Gujarat:
Discrepancies which do not go to the root of the matter and shake the basic version of the witnesses therefore cannot be annexed with undue importance. More so when the all important “probabilities-factor” echoes in favour of the version narrated by the witnesses. The reasons are: (1) By and large a witness cannot be expected to possess a photographic memory and to recall the details of an incident. It is not as if a video tape is replayed on the mental screen; (2) ordinarily it so happens that a witness is overtaken by events. The witness could not have anticipated the occurrence which so often has an element of surprise. The mental faculties therefore cannot be expected to be attuned to absorb the details; (3) The powers of observation differ from person to person. What one may notice, another may not. An object or movement might emboss its image on one person’s mind whereas it might go unnoticed on the part of another; (4) By and large people cannot accurately recall a conversation and reproduce the very words used by them or heard by them. They can only recall the main purport of the conversation. It is unrealistic to expect a witness to be a human tape recorder;Bharwada Bhoginbhai Hirjibhai v. State of Gujarat, (1983) 3 SCC 217, AIR 1983 SC 753, reported headnote.
37This list is written with ordinary observers in mind. It is a statement of general experience about memory, and a cross-examiner should be ready for it. A witness who differs on small things like the colour of a shirt or the exact time of day is not thereby shown to be false, since those are just the details about which memory ordinarily fails. The attack needs to be directed at something that could not reasonably have been forgotten or confused, such as who was present, who struck the blow, or whether the witness was there at all.
38The opposite danger should be kept in mind too. The list is not a shield for every inconsistency, and it does not apply with the same force to a witness who has changed the substance of an account. Nor does it apply well to a professional witness such as an investigating officer, who has notes and who is expected to be exact on matters within the officer’s own duty. The word “root” in the first sentence keeps the rule in its proper place.
Using a prior statement
39The usual means of showing a discrepancy is by a prior statement under section 145 of the Evidence Act and section 162 of the Code of Criminal Procedure. A witness may be contradicted by the statement to the police, and such a contradiction, once proved, affects credit. It does not, of course, make the prior statement evidence of its contents. It is a common error at trial to put a discrepancy to a witness without marking the relevant passage and proving it through the investigating officer. An appellate court cannot rely on a contradiction that has not been established in the proper way, and counsel who wish to rely on the falsity of a part of a witness’s evidence should make sure that this has been done at the right time.
IX. The position today
40The maxim has no application in India as a rule of law or practice. It is at most a rule of caution: a false part of a witness’s testimony means that the rest may be disregarded, not that it must be. A court has a duty to separate the truth from the falsehood and may convict on the residue, even where co-accused are acquitted. That duty ends where the grain and the chaff are so mixed that separation would require a new case to be constructed, or where the substratum of the prosecution case cannot be believed. Evidence falls into the categories of wholly reliable, wholly unreliable and neither, and the court should look for corroboration on material particulars for a witness in the last category. A sterling witness whose word alone may suffice is of very high quality, and the test of such a witness is the truth and consistency of the statement.
History
- Published · 31 October 2026
© 2026 First Precedent Research. Free to share and adapt with credit, for non-commercial use, under CC BY-NC 4.0. A published piece is changed only to correct an error, and each correction is listed under History. If the law changes, we add a linked afterword. Nothing on FPR is legal advice.