The sole issue presented in this case is whether extrinsic evidence of a mistake by a scrivener of a testamentary instrument is admissible in a proceeding to determine the validity of the testamentary instrument. On October 3, 1979, Richards Haskell Emerson, a resident of Lakeville, died at the age of seventy-nine. He had never married and left several cousins as his closest heirs at law. Emerson left a will dated May 19, 1960, and two codicils dated December 4, 1969 (first codicil) and October 24,1975 (second codicil), all of which were offered for probate by the named executor, the Third National Bank of Hampden County, Springfield, Massachusetts.
The Probate Court for the district of Salisbury admitted the will and codicils on February 19,1980, after a hearing on the application for probate. This appeal rises out of the court’s decision to admit portions of the second codicil which the plaintiffs contend were inserted by mistake of the scrivener and therefore do not embody the dispositions intended by the testator.
Contained in the decedent’s will are three articles which set up trusts for the distribution of part of the decedent’s estate and designate, as remainder-men, seven named charitable organizations (1960 *3charities),1 the defendants in this case. In the first codicil to his will, the decedent deleted the seven 1960 charities and substituted another group of eleven charitable organizations (1969 charities) as remaindermen,2 which are the plaintiffs in this ease. There was only one charity named in both the will and the first codicil.3
Subsequently, in 1975, the decedent instructed the trust officer of the Third National Bank of Hampden County to make changes in his will, as amended by the first codicil, so as to qualify the trusts as charitable annuity trusts under the Tax Reform Act of 1969, so that the charitable remainder interests would be allowable as federal estate tax deductions. The trust officer, thereafter, similarly instructed the decedent’s attorney. The attorney, however, in drafting the second codicil, not only made the requested changes but also mistakenly reinstated the 1960 charities, which had originally been named in the will, as beneficiaries under two of *4the three articles of the will which are relevant to this case.4 The decedent, who had never requested or authorized this change, signed the second codicil apparently without realizing the change in beneficiaries.
At the hearing on the application for the admission of the will and the two codicils, the Probate Court heard evidence on the matter and ruled that Connecticut law does not permit the introduction of extrinsic evidence on the issue of mistake, and found that in the absence of such evidence the presumption of the validity of a testamentary instrument mandated the admission of the entire second codicil to probate.5 On appeal to the Superior Court, the defendants moved to strike the plaintiffs’ reasons of appeal. Practice Book § 194. The Superior Court granted the defendants’ motion and held that extrinsic evidence is not admissible to prove that the second codicil, which uses unambiguous language, contained a mistake due to a scrivener’s error and, therefore, should not be admitted to probate. The defendants next joined in the executor’s motion for summary judgment which was *5granted by agreement of the parties.6 From this judgment, the plaintiffs have appealed to this court.
The plaintiffs claim that the lower court erred (1) in failing to find a distinction between a “will construction” proceeding and a proceeding to “admit a will to probate”; (2) in failing to hold that Connecticut cases support the admissibility of extrinsic evidence to prove that material has been mistakenly inserted into a testamentary instrument; (3) in, holding that the second codicil contained no inconsistencies or ambiguities; and (4) in holding that the policy behind the statute of wills is inconsistent with the admissibility of extrinsic evidence to prove lack of testamentary intent for any valid reason including mistake.
I
The plaintiffs’ first argument states that since the issue in this case is the validity of the second codicil and not its meaning, extrinsic evidence should be admissible to prove the scrivener’s error. Specifically, they claim that courts which have considered the question have made a distinction between proceedings to admit a will to probate and will construction proceedings, holding or recognising that extrinsic evidence showing a scrivener’s error is admissible in the former but not in the latter proceeding, absent an ambiguity. See annot., 90 A.L.R.2d 924, 931. Because of the scrivener’s *6error, the plaintiffs claim that the mistake in reinstating the 1960 charities into the second codicil should have been allowed to have been proven by extrinsic evidence and should not have been admitted to probate. The lower court rejected this argument.
“In connection with the distinction between probate or contest proceedings and proceedings to construe a will as affecting the admissibility of extrinsic evidence, it is appropriate to note that in solving the problem of the effect of a draftsman’s mistake, whether upon the validity of the will or upon its operation, the courts must weigh the desirability of effectuating the testator’s true intention and the corresponding undesirability of admitting to probate and putting into operation a will which does not express his true intent, against the danger of invalidating upon parol evidence an instrument executed in compliance with statutory requirements, the danger of relying upon parol evidence being greater in cases of wills than in eases involving other instruments, because the testator’s testimony is not available.” Annot., 90 A.L.R.2d 924, 928, § 2.
In Connecticut, our cases have not, to this point, distinguished between the two types of proceedings. See Stearns v. Stearns, 103 Conn. 213, 130 A. 112 (1925) (will construction proceeding); Comstock v. Hadlyme Ecclesiastical Society, 8 Conn. 254 (1830) (proceeding to establish a will); Avery v. Chappel, 6 Conn. 270 (1826) (proceeding to establish a will). Neither the plaintiffs nor the defendants have presented us with any Connecticut authority which recognizes any distinction between the rules of evidence applicable to either of these two forms of *7proceedings.7 While it is obvious that the purpose behind each type of proceeding may be different,8 we are not inclined to establish a rule which would effectuate such a distinction. It would be unwise to maintain a separate rule regarding the admission of extrinsic evidence for each of these two proceedings. This is because a litigant, knowing that more favorable evidentiary rules await those with claims of mistake due to scrivener’s error, will always strive to phrase his argument in a way so to state such a claim. See generally Stearns v. Stearns, supra, 224; Comstock v. Hadlyme Ecclesiastical Society, supra, 264-66; McFarland v. Chase Manhattan Bank, 32 Conn. Sup. 20, 30, 337 A.2d 1 (1973), aff’d, 168 Conn. 411, 362 A.2d 834 (1975).
This result would elevate the form over the substance of the argument and transform probate proceedings into mere semantic exercises. As an *8example, we point to the situation where a beneficiary under an unambiguous will, which also contains a scrivener’s error, or words which effectuate a different result than that allegedly intended, knows that the extrinsic evidence demonstrating the mistake or a different intention than that expressed in the will is inadmissible in a will construction proceeding. To avoid this substantive barrier, he can merely rephrase his argument to allege that the scrivener did not follow instructions and, therefore, the will, even though otherwise validly executed, should not be admitted to probate since it does not represent the testator’s true intention. See Stearns v. Stearns, supra, 223; see also Travelers Bank & Trust Co. v. Birge, 136 Conn. 21, 27, 68 A.2d 138 (1949). Such a situation would expose many wills to litigation merely because a disappointed beneficiary did not receive exactly what he thought the testator would leave him. If a testator, during his lifetime, represented to a beneficiary that he would receive a certain bequest and, later, the testator changed the bequest in his will but neglected to inform the beneficiary, the beneficiary, knowing that extrinsic evidence is admissible in one but not the other proceeding, could certainly at least allege a scrivener’s error and challenge the will’s admission to probate on the basis that the testamentary document did not represent the true intent of the testator. This would tend to produce needless litigation by transforming a simple will construction proceeding into an admission to probate problem where no problem may have actually existed.
In order to avoid this, we believe that the better course is to recognize that the same evidentiary rules apply to both types of proceedings.
*9II
We now turn to the major issue presented by this case. The trial court held that “parol evidence may not be admitted in the instant case to show that the scrivener erred in drafting the codicil or that the testator mistakenly signed it. Connecticut law does not allow extrinsic evidence of a testator’s intent to be admitted in cases dealing with either will construction or cases challenging the probate of an instrument. While there is an exception to this rule when there is ambiguity on the face of the will or codicil itself, this exception is not applicable in the instant case.” We agree with the trial court.
There is no dispute9 that “[w]hile extrinsic evidence may be admitted to identify the devisee or legatee named, or the property described in a will, also to make clear the doubtful meaning of language used in a will, it is never admissible, however clearly it may indicate the testator’s intention, for the purpose of showing an intention not expressed in the will itself, nor for the purpose of proving a devise or bequest not contained in the will. It is ‘a settled principle, that the construction of a will must be derived from the words of it, and not from extrinsic averment.’ Greene v. Dennis, 6 Conn. 292, 299 *10[1826]." Bryan v. Bigelow, 77 Conn. 604, 614, 60 A. 266 (1905); see Griswold v. First National Bank, 134 Conn. 410, 417, 58 A.2d 256 (1948); Stearns v. Stearns, 103 Conn. 213, 222, 130 A. 112 (1925); Mahoney v. Mahoney, 98 Conn. 525, 537, 120 A. 342 (1923); Jackson v. Alsop, 67 Conn. 249, 252, 34 A. 1106 (1896). That a will is an instrument of gravest importance in the solemn mission of the devolution of the testator’s estate needs no citation. We have said that: “A will is more than a document executed in compliance with the statutory formalities required for'testamentary disposition. ‘A will is the legal declaration of intention as to the disposition of one’s property after death.’ Jacobs v. Button, 79 Conn. 360, 362, 65 A. 150 [1906]; Whitehill v. Halbing, 98 Conn. 21, 23, 118 A. 454 [1922]. ‘ “Will,” as here used [in the statute of wills] . . . means the bequests and devises made by the . . . [testatrix] and expressed in writing and made known through the writing in the manner prescribed by law.’ Hatheway v. Smith, 79 Conn. 506, 511, 65 A. 1058 [1907].” Barnes v. Viering, 152 Conn. 243, 246, 206 A.2d 112 (1964). “We are limited to the language used.” Smuda v. Smuda, 153 Conn. 430, 432, 217 A.2d 59 (1966).
One annotator puts it this way: “What the courts really mean when they say that parol evidence is not admissible to correct a mistake is that where the will as it stands is intelligible and is applicable with. certainty to some person or thing in existence, such evidence is not competent to show the testator’s intention to designate some other person or thing. In other words, the rule against admitting extrinsic evidence to correct mistakes or supply omissions in wills amounts to no more than a slightly different *11statement or application of the general rule against varying, contradicting, or adding to the terms of the will.” Annot., 94 A.L.R. 26, 68-69.
A number of our cases shed light on the question of whether, in the absence of an ambiguity, extrinsic evidence may be admitted to show a scrivener’s error. In an early “admission to probate” case; Comstock v. Hadlyme Ecclesiastical Society, 8 Conn. 254 (1830); the appellants argued that since the scrivener drafted the will contrary to the testatrix’ instructions and since the testatrix executed the will without knowledge of the mistake, the instrument, as drawn, did not represent her will and should not have been admitted to probate. Id., 258. The appellants were the grandchildren and only heirs at law of the testatrix and claimed that the will was void because the scrivener failed to include a legacy of $100 to each of them as the testatrix had directed. The court held that such evidence of a scrivener’s error was not admissible and stated: “The statute, when it required all wills to be in writing, signed by the testator and attested by witnesses, certainly intended, that the evidence, and the whole evidence, of the disposition of property by will, should be the will itself; that the evidence of the intent of the devisor should be derived from the writing, signed by him and solemnly attested; otherwise, innumerable would be the cases, where evidence of mistake would be claimed and proved. To use the language of Ch. J. Best, in the case before cited, some witness would constantly be brought forward to set aside the most solemn instrument. [Provis v. Rowe,] 5 Bing. 435 [1829]. How often is it, that the words used by the scrivener eonvey a different estate from what the testator designed! Yet it has always been decided, that *12parol testimony conld not be admitted to prove, that the devisor meant to give a different estate from what the will expressed. Chappel v. Avery, 6 Conn. Rep. [31,] 34 [1825]. Farrer v. Ayres, 5 Pick. 407 [1827]. It is, says Lord Kenyon, a sacred rule of property not to be departed from. Goodtitle d. Richardson v. Edmonds, 7 Term Rep. [633,] 635, 640 [1798]. Richards v. Dutch & al. 8 Mass. Rep. 506, 515 [1812]. And if it is settled, that you cannot, by parol proof, alter the legal import of the terms used by the scrivener, such a will must either be void, or convey a different estate from the one intended. That such a will is not void, is proved, by repeated declarations of judges, that by the legal construction, they knew that the intent of the testator was frequently violated. [Right v. Sidebotham, 2] Doug. [759,] 763 [1781]. [Williams v. Bosanguet,] 1 Brod. & Bing. [238,] 261 n [1819]. [Denn v. Gaskin, 2] Cowp. [657,] 660 [1777].” Comstock v. Hadlyme Ecclesiastical Society, 8 Conn. 254, 265-66 (1830).
The plaintiffs seek to distinguish this case from Comstock on the basis of the remedy sought. They claim that the remedy sought in Comstock was to rectify the mistake rather than void the will, while the remedy sought in this case is to deny admission to probate. This is an insubstantial distinction. Critical to both remedies sought is the fact that the appellants in both cases attempted to introduce extrinsic evidence to prove the mistake. This, and not the remedy sought, is the key factor in both cases.10 See also McDermott v. Scully, 91 Conn. 45, *1349, 98 A. 350 (1916); Fairfield v. Lawson, 50 Conn. 501, 508, 510 (1883).
It is true that Stearns v. Stearns, 103 Conn. 213, 222, 130 A. 112 (1925), arose in the context of a “will construction” proceeding. That case makes it clear, however, how almost any “will construction” problem can just as easily be argued to involve a question of mistake and, hence, a question of “admitting a will to probate.” Stearns involved the interpretation of the phrase “personal estate” as it appeared in the will of Fannie Nash Stearns. This court explained the various will construction maxims pertaining to the admission of extrinsic evidence in order to construe an ambiguous term in a will. Stearns v. Stearns, supra, 220-22. The court then addressed what appears to be the same argument that is presented in the present case, i.e., that the scrivener failed to draft the will according to instructions and that, therefore, extrinsic evidence of the true purpose and intention of the testator should be admitted to demonstrate such error. The court stated: “There is another and a controlling reason why this evidence was inadmissible. The utmost the offer discloses is the purpose and intention of the testatrix to devise and bequeath all the residue and remainder of her estate to her husband, and her instruction, through her husband acting as her agent, to the serivener to so draft the will, and *14that he failed to draft the will in this particular in accordance with his instructions. The rule is, we believe, universal, that parol evidence of intent cannot be admitted to supply a possible defect or omission in a will occurring through mistake or inadvertence whether of the testatrix or the scrivener. This was declared to he the law of this jurisdiction nearly one hundred years ago and it has remained our law ever since. Avery v. Chappel, 6 Conn. 270 [1826],11 was a hill in equity through which it was attempted to change a provision in a will in which the testator had given the use of his estate during her widowhood to his wife, hy parol evidence that before making his will, the testator declared his intention to give the land in question to his wife until his son Charles should arrive at *15full age. The court found that the scrivener, by mistake, drew the will giving her the estate only during her widowhood. We say of this offer: ‘It is however, urged, that parol evidence ought to be admitted, to correct mistakes in deeds and other writings; and that a court of equity will relieve against such mistakes, as well as against fraud. . . . There is no doubt, that a mistake in a deed or contract may be shown by parol proof; and that, either by a party seeking relief against it in his bill, or setting it up, by way of defense, to rebut an equity. . . . But the plaintiffs in this bill, are obliged to go much further, to obtain the relief sought. They must show, that the testator’s intention may be proved to be different from what appears on the face of the will, by parol evidence that he directed the scrivener so to write it that the wife should enjoy the estate till the son should be twenty-one years of age. It is not believed, that any principle or precedent can be found to establish such a doctrine.’ ” (Emphasis added.) Stearns v. Stearns, 103 Conn. 213, 223-24, 130 A. 112 (1925).
This is similar to that which the plaintiffs request in this case. They seek to introduce evidence that the testator’s true intent was not to divide his estate among the combined eighteen charitable institutions as appears on the face of the will and codicils, but that his true intent, which was allegedly communicated to the scrivener, was to replace the eleven charitable institutions named in the first codicil with the seven charitable institutions originally named in his will. It is clear that whether this is a “will construction” proceeding or a proceeding to “admit a will to probate,”12 where the claimants *16seek to introduce extrinsic evidence of a scrivener’s error resulting in a disposition of property allegedly contrary to the testator’s intention, as expressed to the scrivener, such evidence, absent an ambiguity, is not admissible because “ ‘if such testimony is to be admitted, we do away [with] part at least of the beneficial effect of the statute of frauds, and leave every will exposed to litigation, on a claim of a different intent.’ See also Fairfield v. Lawson, 50 Conn. 501, 509 [1883]; Woodruff v. Migeon, 46 Conn. 236 [1878] ; Hanvy v. Moore, 140 Ga. 691, 696, 79 S.E. 772 [1913]; Lemax v. Lemax, 218 Ill. 629, 75 N.E. 1076 [1905]; Gardner on Wills (2d Ed.) § 105, p. 339.” Stearns v. Stearns, supra, 224-25; see Comstock v. Hadlyme Ecclesiastical Society, 8 Conn. 254, 265-66 (1830);13 Rapp v. Reehling, 124 *17Ind. 36, 23 N.E. 777 (1889). “Claimed errors in the drafting of a will, either by the testator or his scrivener, do not permit the introduction of extrinsic evidence.” McFarland v. Chase Manhattan Bank, N.A., 32 Conn. Sup. 20, 37, 337 A.2d 1 (1973), aff’d, 168 Conn. 411, 362 A.2d 834 (1975). We find that this is the relevant public policy as expressed through our cases; therefore, we will not additionally discuss the plaintiffs’ fourth claim of error set out above.14
We would, however, properly make these observations. It is beyond cavil that “[principles of law which serve one generation well may, by reason of changing conditions, disserve a later one. Dwy v. Connecticut Co., 89 Conn. 74, 99, 92 A. 883 [1915]; Beardsley v. Hartford, 50 Conn. 529, 541 [1883].” Herald Publishing Co. v. Bill, 142 Conn. 53, 62, 111 A.2d 4 (1955). “Experience can and often does demonstrate that a rule, once believed sound, needs *18modification to serve justice better. Ziman v. Whitley, 110 Conn. 108, 114, 147 A. 370 [1929]; Napier v. Peoples Stores Co., 98 Conn. 414, 426, 120 A. 295 [1923].” Herald Publishing Co. v. Bill, supra; see also Funk v. United States, 290 U.S. 371, 54 S. Ct. 212, 78 L. Ed. 369 (1933); MacPherson v. Buick Motor Co., 217 N.Y. 382, 111 N.E. 1050 (1916) (Cardozo, J.). The principle of law, which we reiterate today, has been viable for many years and we have not been given nor do we glean any persuasive reason that justice, experience or logic requires it be changed.
Ill
Finally, the plaintiffs contend that even if the proffered extrinsic evidence would not be admissible to show the scrivener’s error, there exist two inconsistencies and ambiguities in the second codicil which would justify the admission of such evidence under the “ambiguity” exception to which we have previously referred. In claiming that the circumstances of this case make it an “especially appropriate one” in which to admit extrinsic evidence, they also argue that this is so because the evidence of mistake does not come “entirely from outside the will.” The plaintiffs describe the first “ambiguity” as follows: “The last numbered paragraph of the second codicil, labelled ‘III E,’ states the testator’s reason for executing it: ‘It is my intent to create charitable remainder annuity trusts within the meaning of Section 664 (d) (1) and the provisions of these trusts shall be interpreted in accordance with this intent.’ The very next sentence negates any intention to alter the dispositive provisions of the will: ‘In all other respects I do hereby ratify, confirm and republish my said will executed by me *19on May 19, 1960 and my first codicil dated December 4, 1969.’ (Emphasis added.) This statement conflicts with the earlier provisions of the second codicil, which alter the beneficiaries.” (Emphasis in original.)
The plaintiffs claim that it is ambiguous whether the testator intended to change the beneficiaries or whether the specific disclaimer in the last paragraph should be given effect. The trial court rejected this argument on the basis that since the claimed ambiguity did not concern the identity of any beneficiary or the description of the property bequeathed, “[t]he only question is what intention the testator expressed in the will, and absent any latent ambiguity, that is to be determined from the will itself without the use of extrinsic evidence. Morehouse v. Bridgeport City Trust, 137 Conn. 209, 217, 75 A.2d 493 (1950); Mahoney v. Mahoney, [98 Conn. 525, 537, 120 A. 342 (1923)]; McDermott v. Scully, [91 Conn. 45, 48, 98 A. 350 (1916)]; Day v. Webler, 93 Conn. 308, 311, 105 At. 618 (1919); Fairfield v. Lawson, [50 Conn. 501, 510 (1883)]; McFarland v. Chase Manhattan Bank, N.A., [32 Conn. Sup. 20, 27-29, 337 A.2d 1 (1973)].”
We agree with the trial court that no ambiguity is presented by these two sentences in the last article of the second codicil. The provision beginning with “[i.]n all other respects . . .” is obviously separable from the rest of the codicil. Its purpose is clearly intended to demonstrate that all prior testamentary instruments remain unchanged except for those changes outlined in the new testamentary instrument. It creates no ambiguity or inconsistency in the present case.
*20The second “ambiguity” which is claimed to exist arises from the fact that the first codicil substituted the names of the eleven 1969 charities for the seven 1960 charities in three separate articles of the will. The second codicil, however, amended only two of those articles and made no change in the third. As a result, under the second codicil, property will pass to the seven 1960 charities under two articles of the will and to the eleven 1969 charities under a third article of the will. The plaintiffs claim that this result “makes no sense in the testator’s dispositive scheme.”
“The language of the will is controlling. ‘When the purpose of the testator is reasonably clear by reading the words in their natural sense, “we must not be deterred by the conjecture that some remote consequences were not in the testator’s mind, or might not have been quite satisfactory to him, if he had thought of them.” Holmes, J., in Dove v. Johnson, 141 Mass. 287, 290 [5 N.E. 520 (1866)].’ Anderson v. Bean, 220 Mass. 360, 363, 197 N.E. 964 [1915].” Travelers Bank & Trust Co. v. Birge, 136 Conn. 21, 26, 68 A.2d 138 (1949). We cannot say that the testator’s dispositive scheme is ambiguous such that extrinsic evidence should have been admitted to clarify it. The language of the will and codicils is reasonably clear and free from any claimed ambiguity. It “may not seem logical,” as the trial court opined, “to have the 1969 charities and the 1960 charities take under different sections of the will, but there may be different monetary bequests involved, thereby providing the reason.” Nevertheless, the intent of the testator as expressed in the language he used is clear. The question is not what he meant to say, but what is meant by what he did say. See Bryan v. Bigelow, 77 Conn. *21604, 615, 60 A. 266 (1905); Fairfield v. Lawson, 50 Conn. 501, 502 (1883). When the meaning of what is said is clear, “ ‘Why seek by parol to explain that which needs no explanation?’ Hall v. Rand, 8 Conn. [560,] 561, 574 [1831]; Post v. Jackson, [70 Conn. 283, 39 A. 151 (1898)].” Thompson v. Betts, 74 Conn. 576, 579, 51 A. 564 (1902); see Hoenig v. Lubetkin, 137 Conn. 516, 519, 79 A.2d 278 (1951); 1 Locke & Kohn, Conn. Probate Practice § 241, p. 493. “In the absence of compelling reason to the contrary, effect must be given to the testator’s intention as thus expressed and directed in the will. Central Hanover Bank & Trust Co. v. Mason, 129 Conn. 350, 351, 27 A.2d 797 [1942].” Weidlich v. First National Bank & Trust Co., 139 Conn. 652, 657-58, 96 A.2d 547, cert. denied, 346 U.S. 826, 74 S. Ct. 45, 98 L. Ed. 351 (1953).
There is no error.
In this opinion Armentaho and Sponzo, Js., concurred.
The seven 1960 charities are: Sharon Hospital of Sharon, Connecticut; Nantucket Cottage Hospital of Nantucket, Massachusetts; American Society for the Prevention of Cruelty to Animals of New York, New York; and, of Springfield, Massachusetts: Springfield Homo for Aged Men; Springfield Home for Aged Women; Springfield Hospital, and Wesson Maternity Hospital. The Sharon Hospital is not a party to this appeal.
The eleven 1969 charities are: Connecticut Junior Bepublic of Litchfield, Connecticut; American Friends Service Committee, Inc., of Cambridge, Massachusetts; Bobert C. Geer Memorial Hospital of Canaan, Connecticut; Housatonic Psychiatric Center of Lakeville, Connecticut; The Welfare Association of Salisbury, Inc., of Lake-ville, Connecticut; The Salisbury Association of Salisbury, Connecticut; The Scoville Memorial Library of Salisbury, Connecticut; The Connecticut Institute for the Blind of Hartford, Connecticut; The Springfield Museum of Fine Arts of Springfield, Massachusetts; The George Walter Vincent Smith Museum of Springfield, Massachusetts; and Sharon Hospital of Sharon, Connecticut.
The common charity was the Sharon Hospital.
The remaining article relevant to this case continued to provide for the eleven 1969 charities.
The Probate Court made this holding in spite of finding the following: “Substantial and convincing evidence was offered to establish that the testator had directed that the second codicil to his will be prepared only for the purpose of qualifying the charitable bequests under Paragraphs Seventh and Eighth of the will and first eodieil as charitable remainder annuity trusts eligible for an estate tax charitable deduction in the decedent’s estate; and that by scrivener’s error the ‘1960 charities’ were substituted for the ‘1969 charities’ in the 1975 eodieil; and that the proposed 1975 codicil was presented to the testator in a context of innocent misrepresentation of its contents resulting in mistake in the instrument as executed with respect as to which charitable beneficiaries the testator intended to benefit.”
The motion for summary judgment was originally denied because the pleadings were not dosed. See Practice Book § 370. Upon reconsideration, the court ordered entry of summary judgment “by agreement” and found that the will and two codicils had been: duly executed and that the decedent was of sound mind at the time of execution. See Berkeley v. Berkeley, 152 Conn. 398, 207 A.2d 579 (1965).
We do not deem it necessary to undertake any discussion of the matter that jurisdictions outside Connecticut permit extrinsic evidence to prove a mistake in a testamentary instrument.
It is clear that the purpose of a will construction proceeding is to diseem the meaning of the terms of a will, while the purpose behind a proceeding to admit a will to probate is to determine the validity of a will as measured by the requirements of the statute of wills. See General Statutes § 45-161. Chief Justice Arthur Rugg, after noting that “[t]he statutes contain simple, clear and definite rules for the execution and revocation of wills” and that “[r]igorous adherence to these rules will in the end be more likely to effectuate the underlying purposes of those who leave property than an attempt to extract an uncertain intent from equivocal circumstances,” said “[cjourts have no power to reform wills . . . The only means for ascertaining the intent of the testator are the words written and the acts done by him. [Citations omitted.]” Sanderson v. Norcross, 242 Mass. 43, 45-46, 136 N.E. 170 (1922).
Anything that may tend to relax or weaken the effects of the statute of wills is to be guarded against. See Waterbury National Bank v. Waterbury National Bank, 162 Conn. 129, 140, 291 A.2d 737 (1972).
It is also well settled that extrinsic evidence is admissible to prove fraud, incapacity and undue influence. Page v. Phelps, 108 Conn. 572, 143 A. 890 (1928); Wheeler v. Rockett, 91 Conn. 388, 100 A. 13 (1917); Kirby’s Appeal, 91 Conn. 40, 98 A. 349 (1916); Fitzpatrick v. Cullinan, 87 Conn. 579, 89 A. 92 (1913); Lockwood v. Lockwood, 80 Conn. 513, 69 A. 8 (1908); Vivian’s Appeal, 74 Conn. 257, 50 A. 797 (1901); Livingston’s Appeal, 63 Conn. 68, 26 A. 470 (1893); Richmond’s Appeal, 59 Conn. 226, 22 A. 82 (1890); Dale’s Appeal, 57 Conn. 127, 17 A. 757 (1889); Rockwell’s Appeal, 54 Conn. 119, 6 A. 198 (1886); Harrison’s Appeal, 48 Conn. 202 (1880); Canada’s Appeal, 47 Conn. 450 (1880); Mills’s Appeal, 44 Conn. 484 (1877); St. Leger’s Appeal, 34 Conn. 434 (1867).
The dissent states that Comstock v. Hadlyme Ecclesiastical Society, 8 Conn. 254 (1830), “seems” distinguishable from the ease at bar because in Comstock the proponents were seeking to “reform” the will by “adding new provisions to its contents” rather than by *13seeking the “deletion of unintended testamentary bequests, surely a less problematical confrontation with the policy of the Statute of Wills.” (Emphasis in original.) Assuming, arguendo, that deletion and addition each present a “problematical confrontation” of different degree, the policy of the Statute of Wills is nonetheless undermined by the suggestion of such a dichotomy. To permit such evidence, as we hold inadmissible today, would allow others “to make a will for him [the testator], when he sleeps in his grave.” (Emphasis in original.) Avery v. Chappel, 6 Conn. 270, 274 (1826).
In Avery v. Chappel, 6 Conn. 270, 276 (1826) the court stated: “It was decided, by the highest court in South Carolina, after much discussion and deliberation, that parol evidence, even of the person who drew the will, and who was of unimpeachable character, when offered to support the allegation of a mistake in the will, and to prove, that the testator intended to dispose of the property in a manner not apparent oil the face of the will, was not admissible. Rothmaler v. Myers & al. 4 Desaus. 215 [1812], Where there is a complete and plain will in writing, it cannot be altered, or influenced, by parol evidence as to the intention. [Nichols v. Osborn,] 2 P. Wms. [419,] 421 [1727], Evidence as to matter dehors the will, to show the mistake, is insufficient. [Ulrich v. Litchfield,] 2 Atk. [372,] 373 [1742]. Even the instructions for the will are inadmissible, to show a mistake. [Murray v. Jones,] 2 Ves. & Bea. 318 [1813]. [Lansdowne v. Lansdowne,] 1 Mad. Chan. [73,] 81 [1815].” (Emphasis in original.) Id., 276. See McDermott v. Scully, 91 Conn. 45, 49, 98 A. 350 (1916), where this court stated: “It cannot be shown that the testator directed the scrivener to write the will in a form or with a meaning different from that which appears in the will, unless there is a latent ambiguity or equivocation as to the person or subject meant to be described, or unless the evidence is offered to rebut a resulting trust. Avery v. Chappel, 6 Conn. 270, 275 [1826]. The rule and the exception, and the reasons underlying them, have been long since stated by our court. Their application still affords matter for controversy.”
In fact, it is clear that if Stearns v. Stearns, 103 Conn. 213, 130 A. 112 (1925), were solely a “will construction” case, thus allegedly *16limiting its precedential value, this argument could not even have been made. In a will construction proceeding, a claimant cannot allege that the testamentary instrument is not the testator’s will; he can only demonstrate an ambiguity or other construction problem appearing within the four comers of the will.
Our holding today is limited to the case where a mistake due to a scrivener’s error with no resulting ambiguity is alleged. It is clear that there are certain types of “mistakes” which would merit the introduction of extrinsie evidence as proof thereof. In In re Gluckman’s Will, 87 N.J. Eq. 638, 641, 101 A. 295 (1917), the court stated: “Where a testator, in addition to complete testamentary mental capacity, is in full enjoyment of average physical and educational faculties, it would seem that in the absence of fraud or of undue influence, a mistake, in order to defeat probate of his entire will, must in substance or effect really amount to one of identity of the instrument executed; as, for instance, where two sisters, in one'ease, or a husband and wife, in another, prepared their respective wills for simultaneous execution, and through pure error one executed the other’s, and viee versa. Anon., 14 Jur. 402 [1850]; Re Hunt, L.R. 3 P. & D. 250 [1875]; Nelson v. McDonald, 61 Hun. 406, 16 N.Y. Supp. 273 [1891], Short of this, however, or of something amounting in effect to the same thing, it is against sound public policy to permit a pure mistake to defeat the duly solemnized and completely competent testamentary act. It is more important that the probate of the wills of dead people be effectively *17shielded from the attacks of a multitude of fictitious mistakes than that it be purged of wills containing a few real ones. The latter a testator may, by due care, avoid in his lifetime. Against the former he would be helpless.” See 1 Page, Wills (Bowe-Parker Rev.) §§ 13.6-13.13; see also Shulman v. Shulman, 150 Conn. 651, 664, 193 A.2d 525 (1963); Sansona v. Laraia, 88 Conn. 136, 138, 90 A. 28 (1914). Consistent with the policy of the Statute of Wills as well as our repeated rule that the testator’s intent is to be gleaned from the words that in fact appear in his will, the category of mistake referred to in Rockwell’s Appeal, 54 Conn. 119, 6 A. 198 (1888), partakes of that type of mistake that may servo to defeat the admission of a will to probate as is sot out in In re Gluckman’s Will, 87 N.J. Eq. 638, 641, 101 A. 295 (1917). See In re Snide, 52 N.Y.2d 193, 418 N.E.2d 656 (1981).
In their reply brief the plaintiffs argue that “[w]hether or not the second codicil was read is not conclusive and is an issue of fact properly not before this court,” claiming, inter alia, that such an issue is basically not within the truth of the allegations of the reasons of appeal concededly admitted for purposes of the motion to strike. In our disposition of this appeal, wo need not even reach this issue.