This is a proceeding for an accounting involving the construction of the will of Alessandro Fabbri, who died on February 6, 1922, a resident of Bar Harbor, Maine.
The will was executed on April 13, 1920. At that time the testator was an unmarried man with two brothers and four sisters. One of his brothers and all of his sisters were residents of Florence, Italy. A brother, Ernesto, his wife, Edith, and testator were residents of the State of Maine. In his will, after providing for the payment of all debts, the testator made a bequest of $20,000 to Ernesto and created a trust for the benefit of one Charles Cushing, a friend. Cushing was to receive the income of the fund up to $2,400 a year. Annual income in excess of that amount and the entire principal, upon Cushing’s death, were made parts of the residuum of the estate. The residue was to be divided into two separate trust funds, income payable to Ernesto and Edith during their respective lives. Upon the death of Ernesto the principal of the trust created for his benefit was to be distributed among three of testator’s sisters in Florence or to their surviving children if any of the sisters predeceased the testator. Upon Edith’s death the principal of her trust was to be distributed pursuant to article Sixth of the will, which reads as follows: ‘ ‘ Sixth * * * upon the death of my brother’s wife, Edith S. Fabbri, either before or after me, the principal of the trust fund created for her benefit, shall be distributed to and among such issue of Teresa F. Clark, daughter of the said Edith S. Fabbri, as I may designate in writing.” Edith died on December 17, 1954. At that time her daughter, Teresa F. Clark, had two children, James Cameron Clark, Jr., and Edith Clark Milliken, the appellants here. At no time after *239the execution of the will did testator make any further designation as to the remainder of the trust.
Respondents, representatives of testator’s heirs at law and next of kin, contend that the provision under consideration unmistakably reveals that the testator did not intend to dispose of the remainder interest of the trust by his will but merely indicated that such disposition might be accomplished by a subsequent additional instrument. They conclude that since no additional instrument was ever executed, the entire remainder must pass by the rules of intestate succession.
The trustee brought this proceeding to determine whether the principal of the trust was to be divided between the issue of Teresa Clark or was to pass, as in intestacy, to representatives of the testator’s next of kin.
At the time of his death testator was domiciled in Maine and, since no contrary intent appears in the will, the law of that State determines the validity and effect of the will with respect to the disposition of personal property (1 Davids on New York Law of Wills, § 452; Decedent Estate Law, § 47; Matter of Gifford, 279 N. Y. 470, 474-475). As to that portion of the trust assets consisting of an interest in real property situated in New York, the laws of New York State apply. (Decedent Estate Law, § 47.) As is conceded by the respondent, however, as far as the questions on this appeal are concerned, there is no conflict between the law of Maine and that of New York.
It can be stated at the outset that decisions such as Smith v. Floyd (140 N. Y. 337), cited by the appellants, deal exclusively with imperative powers of appointment and have no pertinency to this case. In the case at hand, the testator did not grant a power to a third person by which a future estate might be created. Since a will is inoperative and wholly ineffective until the death of the testator, after the execution of this will, the testator’s authority to dispose of all of his property remained unfettered (Towler v. Towler, 142 N. Y. 371, 374-375).
There remains for our consideration the sole question: who is entitled to receive the principal of the trust for Edith A. Fabbri?
The prime consideration here as in all construction proceedings is the intention of the testator as expressed in the will. All rules of interpretation are subordinated to the requirement *240that the actual purpose of the testator be sought and effectuated as far as is consonant with principles of law and public policy.
This intent, as we have often said, must be gleaned not from a single word or phrase but from a sympathetic reading of the will as an entirety and in view of all the facts and circumstances under which the provisions of the will were framed. (Collister v. Fassitt, 163 N. Y. 281; Robinson v. Martin, 200 N. Y. 159, 164; Williams v. Jones, 166 N. Y. 522, 533; March v. March, 186 N. Y. 99, 103.)
As this court recently stated in Spencer v. Childs (1 N Y 2d 103, 106-107) which also involved a problem of construction: “ Cases such as the present and Matter of Daly [1 N Y 2d 100] — also decided today * * * — well illustrate the aptness of Judge Learned Hand’s wise and trenchant observation that courts should be wary of making ‘ a fortress out of the dictionary ’, since there ‘ is no more likely way to misapprehend the meaning of language * * * than to read the words literally, forgetting the object which the document as a whole ’ seeks to achieve. (Cabell v. Markham, 148 F. 2d 737, 739; Central Hanover Bank & Trust Co. v. Commissioner, 159 F. 2d 167, 169.) ‘ A word * * * may vary greatly in color and content ’ according to the intent of its author and the circumstances under which it is used. (Towne v. Eisner, 245 U. S. 418, 425.). ”
If the court upon reading the will in this setting discerns a dominant purpose or plan of distribution, the individual parts of the will must be read in relation to that purpose and given effect accordingly (Roe v. Vingut, 117 N. Y. 204, 212). This is true despite the fact that a literal reading of the portion under construction might yield an inconsistent or contradictory meaning because of the use of awkard language inadvertently or carelessly chosen (Haug v. Schumacher, 166 N. Y. 506, 513; Williams v. Jones, 166 N. Y. 522, 533, supra). As this court pointed out in an early opinion: “ If we can see that the inapt, or careless, use of language by the testator has created the difficulty in ascertaining his intention, but, nevertheless, feel certain as to what he meant, from reading the whole instrument in connection with the clause in question, we may subordinate the language to that meaning*. ’ ’ (Matter of Miner, 146 N. Y. 121, 130-131.) By application of these principles to the language before us, we have concluded that it was the testator’s purpose to make a gift of the remainder of the Edith Fabbri trust to the issue of Teresa Clark. As we read the particular *241language and the will, the nomination of that class as the group entitled to the remainder was unconditionally made; only a designation by way of apportionment was reserved for future determination.
The will reveals a general plan or design to distribute thoroughly and completely Alessandro’s entire estate. His brother, Ernesto, and Ernesto’s wife, Edith, were to be the recipients of the principal gift. Upon their deaths half of the residuary estate was to pass to the three married sisters in Florence, Italy, members of the Italian branch of the family, and the other half to the posterity of Ernesto and Edith, the American branch. The remaining brother, Egisto, and an unmarried sister, both of Florence, were not mentioned. As evidenced by the will, Ernesto and his wife Edith were the foremost objects of the testator’s bounty. It is no more than natural to assume that their progeny were held in high favor by Alessandro, who was himself unmarried, and that they too would be natural objects of his generosity when he made his testamentary disposition. Although Teresa, the only child of Edith and Ernesto, received no direct gift in the will, testator undoubtedly considered that she would benefit by the substantial gifts to her parents. As to Teresa’s issue, however, the grandchildren of Edith and Ernesto, testator made specific provision by designating them recipients of the remainder of the Edith Fabbri trust. The gift was accomplished in article Sixth of the will. The wording “ such ” and “ as I may designate ”, upon which respondents place such great stress, is not inconsistent with the gift. That language indicates only the reservation of a right to prefer certain members of the issue. The reservation is understandable when viewed in conjunction with the membership of the Clark family at the time of the making of the will. When the will was written Teresa Clark had but one child and he was then under two years of age. Under ordinary circumstances this class could be expected to increase during testator’s lifetime. It was entirely possible that over the years some of the grandnephews and grandnieces might assume different places in the affection of the testator. Aware of this possibility, testator desired to allow their personal relationships to develop while at the same time preserving the inheritance for the grandnephews and grandnieces as a class.
By the language of article Sixth, therefore, testator indicated a paramount desire that the remainder of the Edith Fabbri *242trust pass to and be enjoyed by the issue of Edith’s only child, Teresa. Had testator so desired, a further designation might have been executed, stipulating that the issue take in varying amounts. However, no such designation was made and consequently the issue take equal portions.
It has been argued that such a determination constitutes a redrafting of the will by the court, since testator himself never stipulated the amounts in which the issue should take. The argument is without merit. The issue take the remainder in equal shares, not because we so determine, but by virtue of the absence of an apportionment on the part of the testator. As we read the will, apportionment or allocation was not a condition precedent to the gift and, consequently, testator’s failure to exercise his prerogative of allocating the gift in different proportions constitutes no obstruction to the fulfillment of his paramount intent to restrict the use and benefit of the remainder to the issue of Teresa F. Clark. Since the dominant intent was inferable from the instrument itself, its fulfillment constitutes no redrafting of the will.
It appears to us that the construction we have placed upon the language in question more nearly comports with the intention of the testator as manifest in the will than does the interpretation urged by the respondents. The suggested construction ignores the admonition that “ Qui haeret in litera, haeret in eortice— [He who considers merely the letter of an instrument goes but skin deep into its meaning] ”. (Riggs v. Palmer, 115 N. Y. 506, 510.)
Respondents’ argument that the language is clear and free from any ambiguity, as they interpret it, would compel a conclusion on our part that the testator deliberately did not dispose of a substantial portion of his estate and intended that, in the event of his death, without making a further writing, the property should pass by the laws of intestacy. It would be tantamount to saying that he intended by the language employed to indicate a temporary disinheritance of the grandchildren of Ernesto and Edith, the brother and sister-in-law between whom he had divided the income of his residuary estate and who were apparently his favorite relatives, in favor of relatives for whom he had already provided, or relatives he had not mentioned in his will. Such a result contradicts not only the testamentary scheme and dominant purpose discernible in the *243will but runs counter, also, to experience. A testator, by tbe act of tbe making of a will, easts grave doubt on any assumption that be expressly intends to cbance dying intestate as to any portion of bis property. Indeed, tbe law bas taken cognizance of tbis teaching of common experience and crystallized it into a presumption, expressed in various terms, but requiring essentially that tbe courts favor a construction wbicb avoids partial intestacy and adopt one wbicb results in a complete disposition of tbe estate (Haug v. Schumacher, 166 N. Y. 506, 514-515, supra; Lewis v. Howe, 174 N. Y. 340, 346). Tbis presumption against intestacy is particularly weighty where tbe gift, as here, is made out of tbe residuary estate (Matter of Hayes, 263 N. Y. 219, 225). Tbe opinion in tbe Hayes case (p. 225) quotes tbe following forceful statement of tbe rule: ‘ ‘ ‘ Tbe idea of anyone deliberately purposing to die testate as to a portion of bis estate, and intestate as to another portion, is so unusual, in tbe history of testamentary dispositions, as to justify almost any construction to escape from it.’ (2 Redfield on Wills [3d ed.], p. 235.) ”
It may be that tbe wording employed by tbe draftsmen in article Sixth was inappropriate or ineptly arranged. It may also be true that when tbis language is segregated from tbe rest of tbe will and subjected to a literal scrutiny, it may be susceptible of a meaning different from tbe one arrived at here. But as pointed out above tbe process of interpretation is not so rigid. Tbe search for intention is difficult enough without tbe imposition of undue and obstructive requirements that would confine tbe court’s investigation to a mere reading of tbe language before it. Such an approach would constitute an abdication of tbe court’s duty to accomplish, as nearly as possible, tbe disposition actually intended by tbe testator, and would reduce its task to a simple exercise in semantics. It would overlook tbe very real possibility of distortion and lack of conformity between tbe disposition actually intended and tbe one indicated by an overly strict and pedantic reading of tbe language employed. Tbe law takes into consideration tbe relative inadequacy of words as a vehicle for communicating intent and seeks tbe purpose of tbe testator by means of a more thorough and realistic approach. As already indicated, by reading tbe language in light of tbe rest of tbe instrument, tbe circumstances surrounding its formulation and tbe inferences supplied by com*244mon experience (in this case the presumption against intestacy), the court seeks to minimize the possibility that testator’s true purpose will be frustrated by an unwarranted dependence on an ill-chosen word or phrase.
This approach has it limits. There are cases where the language employed is so clear and unmistakable as to convey only one meaning even when read in its proper setting. In such cases the intent evidenced by the language is given effect without reference to external circumstances or rules of construction. In these instances only one meaning is reasonable, and a variance from the language would constitute not an interpretation but a refabrication of the will by the court (see Matter of Watson, 262 N. Y. 284, 294). Such is clearly not the case here. It can hardly be said that the import which respondents place on the clause in question flows unalterably and unmistakably from a reading of the language in conjunction with the rest of the will. Even without seeking to harmonize the clause with the testamentary scheme otherwise apparent, when the language is given a plain reading, it is susceptible of both the constructions contended for here. Under such circumstances it is the court’s duty to employ, as was done in the instant case, every means at its disposal to ferret out and effectuate the actual intent of the testator.