dissenting.
The only vestiges of a parenting relationship between Mr. Matson and his daughter Shannon is that he lives in the same small community and that he visits Shannon for an hour each day when he is in town. On these minimal facts the majority conclude that this relationship is sufficient to constitute Mr. Matson’s “custody” of Shannon and that, therefore, the less demanding “best interests” standard is applicable. With these conclusions, I cannot agree.
While the contours of “custody” are admittedly not amenable to precise definition,1 for that concept to be any more than a hollow legal label requires a more searching inquiry into the indicia of a parenting relationship. In the present case, I think the conclusion ineludible that Mr. Matson *301bas abdicated all parental authority to the Christensens, and that it is the Christensens who in fact seek “custody” of Shannon. It is uncontroverted that the father and the Christensens intend for Shannon to permanently reside in the Christensens’ home. The Christensens provide all financial support for Shannon. Most importantly, there is absolutely no evidence that Mr. Matson makes any decisions regarding the rearing of Shannon, except of course the “decision” to authorize the Christensens to raise his daughter.
One ramification of the majority’s unwarranted characterization of this dispute as between two parents is that the resultant award of custody to Mr. Matson authorizes him to unilaterally remove the child from the Christensen home. The father has a second wife and baby daughter, yet there is no evidence regarding Shannon’s relationship with either her stepmother or stepsister. There is no evidence of the father’s or stepmother’s current parenting abilities, neither of whom were interviewed by the custody investigator. Indeed, there is no evidence regarding even the physical condition of the Matson home. That it may be possible to modify the decree were Mr. Mat-son to remove Shannon from the Christensen home does not, in my view, sufficiently safeguard the welfare of the child.2
Of more general concern, the unduly expansive characterization of “custody” adopted by the majority can only invite future instances of subterfuge. A third party need only elicit one of the parents to nominally seek custody in order to substantially increase the likelihood that the other parent’s claim for custody will be defeated.3
Further, I note that in disputes between a parent and a non-parent this court has held that absent a finding of parental unfitness or abandonment, the non-parent must show by a preponderance of the evidence that an award of custody would be “clearly detrimental” to the child. Britt v. Britt, 567 P.2d 308, 310 (Alaska 1977); Turner v. Pannick, 540 P.2d 1051, 1055 (Alaska 1975). The majority’s treatment of this case substantially undermines the “clearly detrimental” standard, a rule reflective of the historic preference for parental custody. Were the majority of the court to conclude that the preference for parental custody should only be expressed in terms of the “best interests” of the child, I think it far preferable to directly confront the rule of Turner v. Pannick than to affirm it on the one hand, while so severely limiting its effect on the other that it has little continuing vitality.4
T conclude, therefore, that this dispute is best characterized as one between a parent and a non-parent. Accordingly, I view the dispositive question as whether sufficient evidence was introduced to demonstrate that an award of custody to the mother would be clearly detrimental to Shannon.
The clearly detrimental standard comprehends those situations where an otherwise fit parent is denied custody because compelling considerations militate in favor of an award of custody to a non-parent. It is erroneous to interpret the standard in a fashion which renders it indistinguishable *302from a showing of parental unfitness.5 I add, further, that it is improper to infer that the clearly detrimental standard requires the same quantum of evidence in all cases where a non-parent seeks custody over the objection of a fit, natural parent. In other contexts, this court has noted that it is undesirable to resolve sensitive custody disputes by the “mechanistic application” of rigid rules or presumptions. E.g., Johnson v. Johnson, 564 P.2d 71, 74 (Alaska 1977) (rejection of “tender years” doctrine). The preference for awarding custody to natural parents should similarly be adaptable to the many different factual contexts in which custody disputes arise. See Ellerbe v. Hooks, 490 Pa. 363, 416 A.2d 512, 515 (1980) (concurring opinion). Thus, for example, where a non-parent is the child’s “psychological parent,” or where the child has lived with the non-parent for a substantial length without contact with the natural parent, the parental preference should be accorded less weight.
Mindful of the broad discretion necessarily delegated to the trier of fact, I nonetheless conclude that the special master was clearly erroneous in finding that it would not be clearly detrimental to sever Shannon’s ties with the Christensen family. Shannon, now seven years old, has lived in the Christensen home since February 1976. She identifies Mrs. Christensen as her “psychological parent.” In addition, she enjoys close relationships with her “siblings,” the Christensens’ natural children. The court-appointed custody investigator concluded that “to remove the child from the setting she is now in, [sic] would be extremely disruptive to her and would not be in her best interest.”
The contrasting consideration was that the mother appears to offer an adequate home environment. But, the child had had little contact with her mother during the three years before the hearing, and no contact with her stepfather. The master nonetheless concluded that Shannon could adjust to the new environment without great difficulty, or at least without lasting effect. The only “expert” evidence available to the master, the report of the custody investigator, concluded otherwise.6 In addition, it is noteworthy that a guardian ad litem was not present at the hearing. See Carle v. Carle, 503 P.2d 1050, 1054 (Alaska 1972).
In my view, therefore, it was clearly erroneous for the master to find that an award of custody to the parent would not be clearly detrimental to Shannon. Here, stability itself may be the most salient consideration. See Gruenberg & Mackey, A New Direction For Child Custody in Alaska, 6 U.C.L.A.Alaska L.Rev. 34 (1976). To remove Shannon from the only stable home environment she has known would sever her bond to her “psychological parent,” see Hoy v. Willis, 165 N.J.Super. 265, 398 A.2d 109 (1978), and to her “siblings.” Nichols v. Nichols, 516 P.2d 732 (Alaska 1973). Such considerations are sufficient, in my view, to establish a showing of a clear detriment. Contra, Hickey v. Bell, 391 P.2d 447 (Alaska 1964).
In sum, I would reverse the judgment of the superior court and would remand this case with instructions to award custody of Shannon to the Christensens.
. Delgado v. Fawcett, 515 P.2d 710, 712 (Alaska 1973).
. Since the court rests its award of custody on such attenuated grounds, it is surely not idle conjecture to consider whether such future changes as the father’s proximity, or even his, desire to visit Shannon on a regular basis could serve as the basis for an action to modify the custody order.
. In the present case, it is uncontroverted that the Christensens have financed Mr. Matson’s litigation expenses.
. Several justices of this court have previously opined that little clarity is obtained in attempting to articulate a meaningful distinction between the “best interests” and the “clearly detrimental” standards. See Turner v. Pannick, 540 P.2d at 1056 (Rabinowitz, C.J., concurring).
Even assuming that such a meaningful distinction can be drawn, the Turner rule is problematic because it requires a categorization of whether it is a parent or a non-parent who seeks “custody”. As the present case cogently illustrates, the concept of “custody” is not amenable to such a formalistic inquiry. It is thus arguably undesirable for so much to revolve about the label which a trial court attaches to a parent’s proposed relationship with the child.
. The Britt case inadvertently invited such a restrictive interpretation. We stated:
The trial court’s findings that Gale Britt was an unfit parent (or in the terms of the Turner opinion, that placing custody with her would be clearly detrimental to the child)
567 P.2d at 310.
. I am constrained to add that the present case evinces the desirability of procuring the testimony of an expert, such as a child psychologist. See Starkweather v. Curritt, 636 P.2d 1181, 1183 n.3 (Alaska Nov. 27, 1981).