In this appeal, we review the circuit judge’s grant of summary judgment in favor of appellee First Unum Life Insurance Company. We hold that summary judgment was improper because genuine issues of material fact remain to be determined. Therefore, we reverse and remand.
On January 1, 1994, First Unum issued a disability insurance policy under which appellant James Elam was an insured. The policy provided for payment of benefits to age sixty-five in the event an insured became disabled. On August 8, 1994, First Unum began paying benefits to Elam, who was then age forty-three, as a result of Elam’s bipolar disorder. However, benefits were terminated *56after twenty-four months on the basis of the following policy limitation:
MENTAL ILLNESS LIMITATION
Benefits for disability' due to mental illness will not exceed 24 months of monthly benefit payments unless the insured meets one of these situations: •
[situations,not applicable],
‘Mental illness’ means mental, nervous or emotional diseases or .disorders of any type. . .
On May 9,. 1997, Élam filed suit against First Unum seeking a declaration that he was entitled to further benefits because his bipolar disorder did not fall within the policy’s mental illness limitation. He later filed a motion for summary judgment, arguing that the mental illness limitation was not triggered because his bipolar disorder was biological in origin. Attached to his motion were the affidavits of two' medical doctors, Bradley C. Diner and Charles Bowden. Dr. Diner stated in his affidavit that bipolar disorder is a biological condition with hereditary predisposition and that an alteration in brain chemistry is responsible for the mood disturbances and altered thought processes that accompany the disorder. Dr. Bowden stated in his affidavit that “there is no longer any reasonable doubt among informed members of the- medical, community that Bipolar Affective Disorder has a biological.origin....”-
First Unum responded with its own motion for summary judgment, arguing that the policy "unámbiguoúsly excluded further benefits for disability due' to bipolar disorder. Attached to First Unum’s motion were excerpt’s from the deposition of Dr. Diner and excerpts-from the. deposition of Dr. Joe Backus, Elam’s. treating psychiatrist.-These depositions were later admitted in their entirety through Elam’s response to First Unum’s motion. They revealed the following pertinent information: ...
1. Bipolar cüsordér is a mental disease which reflects mood swings of mania and depression. It is diagnosed based upon behavior, clinical presentation, psychological testing, and patient history.
2. There is no-treatment for bipolar disorder other than drugs. Therapy is used to help educate the patient about the disease and to help him adapt to it.
*573. There are no specific diagnostic markers for any mental disorder. They cannot be detected with a brain scan or a blood test.
4. Bipolar disorder is a biologically-based mood disorder. However, the psychiatric community recognizes bipolar disorder as a mental illness. It is typically treated by psychiatrists.
Elam’s response to First Unum’s motion also contained a supplemental affidavit from Dr. Diner in which the doctor quoted the following from the Diagnostic and Statistical Manual of Mental Disorders:
Although this volume is titled The Diagnostic and Statistic [sic] Manual of Mental Disorders, the term mental disorder unfortunately implies a distinction between ‘mental’ disorders and ‘physical’ disorders that is a reductionistic anachronism of mind/body dualism. [CJompelling literature documents that there is much ‘physical’ in ‘mental’ disorders and [much] ‘mental’ in ‘physical’ disorders. The problem raised by the term ‘mental’ disorders has been much clearer than its solution, and, unfortunately, the term persists in the tide of DSM-IV because we have not found an appropriate substitute.
Finally, Elam filed his own affidavit to which he appended numerous articles from the lay press discussing the biological basis of what are traditionally perceived as mental disorders.
Following a hearing, the circuit judge issued an order granting First Unum’s motion for summary judgment. He found that the “common, ordinary and lay understanding” of the term “mental illness” as used in the policy encompassed bipolar disorder. Thus, he declared the term unambiguous, and First Ünum’s mental illness limitation was upheld. Elam appeals from that order.
In reviewing summary-judgment cases, we determine whether the trial court’s grant of summary judgment was appropriate based on whether the evidence presented by the moving party left a material question of fact unanswered. Norris v. State Farm Fire & Cas. Co., 341 Ark. 360, 16 S.W.3d 242 (2000). In cases involving interpretation of a clause in an insurance contract, the construction and legal effect of the contract are questions of law unless the meaning of the contract depends on disputed extrinsic evidence. Smith v. Prudential Prop. & Cas. Co., 340 Ark. 335, 10 S.W.3d 846 (2000). An insurance policy is to be construed in its plain, ordinary, and popular sense. Norris v. State Farm, supra.
*58We begin our analysis by discussing the case that touches most closely upon the issue at hand. In Arkansas Blue Cross & Blue Shield v. Doe, 22 Ark. App. 89, 733 S.W.2d 429 (1987), Doe was an insured under a group health policy issued by Blue Cross. After she was diagnosed with bipolar disorder, Blue Cross limited her benefits on the basis of a policy provision that restricted coverage for mental, nervous, or psychiatric conditions. Doe filed suit seeking coverage as if her condition were a physical illness. She presented evidence, mostly in the form of medical testimony, that her illness should be classified by cause rather than by symptoms and that the cause of her illness was biological. Blue Cross presented evidence that bipolar disorder should be classified as a mental disorder. The circuit judge, sitting as fact-finder, determined on the conflicting evidence that Doe’s condition was physical rather than mental and thus ruled that Blue Cross’s coverage restriction did not apply to Doe’s bipolar disorder. On appeal, we upheld the judge’s finding on the basis that it was not clearly against the preponderance of the evidence, and we gave due regard to his superior ability to judge the credibility of the witnesses.
Appellant argues that Doe stands for the proposition that, as a matter of law, a mental illness coverage limitation is not triggered if the insured’s disorder is biological in origin. Appellant reads too much into our holding in that case. Our affirmance of the trial judge’s decision in Doe was made out of deference to his factual findings, based upon the conflicting evidence before him. We did not hold that, as a matter of law, the insured’s bipolar disorder was biological in nature and therefore not subject to the coverage limitation. Therefore, Doe has a limited application in the case at bar. The issue before us is not whether the evidence supports the trial judge’s finding or whether that finding is clearly erroneous but whether the mental illness coverage limitation in First Unum’s policy is ambiguous.
Numerous courts have struggled with the issue of whether coverage limitations for mental illness are ambiguous. Some courts have held that policy language that fails to define “mental illness” in a very specific way is inherendy ambiguous. See Lang v. Long Term-Disability Plan, 125 F.3d 794 (9th Cir. 1997) (depression associated with fibromyalgia); Mongeluzo v. Baxter Travenol Long Term Disability Benefit Plan, 46 F.3d 938 (9th Cir. 1995) (anxiety and depression); Patterson v. Hughes Aircraft Co., 11 F.3d 948 (9th Cir. 1993) (head*59aches and depression); Phillips v. Lincoln Nat’l Life Ins. Co., 978 F.2d 302 (7th Cir. 1992) (encephelopathy); Kunin v. Benefit Trust Life Ins. Co., 910 F.2d 534 (9th Cir.), cert. denied, 498 U.S. 1013 (1990) (autism); Dorsk v. Unum Life Ins. Co., 8 F. Supp. 2d 19 (D. Maine 1998) (obsessive compulsive disorder); Gareis v. Benefit Ass’n of Railway Employees Ins. Co., 284 Minn. 262, 169 N.W.2d 730 (1969) (nervous system disorder). These courts have tended to view the cause of an illness as a critical consideration in determining whether the illness was mental or physical. Other courts have held that, if the insured’s symptoms, as viewed by a lay person, indicate the presence of a mental illness, the coverage limitation is unambiguous. See Prudential Ins. Co. v. Doe, 140 F.3d 786 (8th Cir. 1998) (recurrent major affective disorder); Lynd v. Reliance Standard Life Ins. Co., 94 F.3d 979 (5th Cir. 1996) (major depressive disorder); Stauch v. UnisysCorp., 24 F.3d 1054 (8th Cir. 1994) (depression and fatigue); Brewer v. Lincoln Nat’l Life Ins. Co., 921 F.2d 150 (8th Cir. 1990), cert. denied, 501 U.S. 1238 (1991) (affective mood disorder); Equitable Life Assur. Soc. v. Berry, 212 Cal. App. 3d 832, 260 Cal. Rptr. 819 (1989) (bipolar disorder). Yet, other cases have focused on the treatment given to the insured. If the treatment is of the type associated with a mental illness, the policy limitation applies. See Blake v. Unionmutual Stock Life Ins. Co., 906 F.2d 1525 (11th Cir. 1990) (postpartum depression). If the treatment is of a type associated with physical illness, the policy limitation does not apply. See Simons v. Blue Cross & Blue Shield of Greater New York, 144 A.D.2d 28, 536 N.Y.S.2d 431 (1989) (anorexia nervosa).
The evidence presented below by both parties in support of their motions for summary judgment reflects, just as the above-cited cases do, the variation in thought that exists as to whether an illness should be classified as mental or physical. Doctors Diner and Backus stated their belief that bipolar disorder has a biological origin. Flowever, Dr. Diner acknowledged that it is a mental disease and a mood disorder, typically treated by psychiatrists. Dr. Backus acknowledged that the definition of mental illness might be broad enough to include behavioral problems that are neurological or chemical in origin. The Diagnostic and Statistical Manual, while it questions the use of the term “mental disorder” also recognizes that no appropriate substitute has been found for the term.
The differing approaches taken in the above-cited cases and the conflicting thoughts voiced by the medical experts herein illus*60trate the difficulty inherent in attempting to declare a particular disorder either a mental illness or a physical illness as a matter of law. When we are interpreting the term “mental illness” as it is used in health and disability insurance plans, we are dealing with the complexity of the human condition. The number and type of disorders and diseases that may be visited upon human beings are as many and varied as the humans themselves. We are not so simple or dualistic that our minds and bodies work, or disfunction, separately. The manner in which we become sick, the symptoms we exhibit, and the manner in which we are healed often involve the mind, the body, or a combination of both. Further, advances in medicine and the wide dissemination of medical knowledge among the lay public has had the effect of altering perceptions as to what constitutes a mental illness. Thirty-seven years ago, our supreme court was presented with a case in which an insurer denied benefits to an insured with arteriosclerosis on the basis that he was suffering from a mental infirmity. See Mutual Benefit Health & Acc. Ass’n v. Rowell, 236 Ark. 771, 368 S.W.2d 272 (1963). We venture to say that, today, it would be beyond question that the insured in that case did not suffer from a mental illness. Perhaps, thirty-seven years from now, an illness that is generally perceived to be a mental illness may likewise be thought of as unquestionably physical.
The question of what constitutes a mental illness is obviously a mutable, evolving concept. We are therefore reluctant to adopt a hard and fast rule, as the Eighth Circuit did in Brewer, supra, when it declared that the nature of an illness is determined by its symptoms. Illnesses can seldom be classified so simply. What may be manifested as a mental illness may in fact be a physical, biological disorder. For example, a person suffering from a brain tumor may exhibit the type of behavior normally associated with a severe mental illness. However, there are few persons who would agree, upon learning of the existence of the tumor, that the person was suffering from a mental illness. Alternatively, what may be manifested as a physical illness may in fact be a mental illness. Stomach disorders or headaches may be symptoms of stress or depression. We are likewise wary of adopting a rule that cause alone should determine whether an illness is mental or physical in nature. As Drs. Diner and Backus noted below, chemical or hormone imbalances may cause depression or bipolar disorder. It is also possible that mental disfunction may cause physical disability. For example, paralysis, fatigue, or *61other physical conditions may be caused by mental difficulties such as anxiety. Likewise, the type of treatment accorded is seldom conclusive in determining the nature of the illness. A person with a chronic physical condition such as fibromyalgia, Parkinson’s disease, autonomic disfunction, or chronic fatigue syndrome may be treated with psychotherapy to help cope with their disease.
The point of the above discussion is to express our conclusion that the term “mental illness” as it is used in First Unum’s policy, is susceptible to any number of interpretations. Neither the legal authorities, the medical authorities, nor the public can agree on a particular definition of the term. It may mean, for example, mental illness in the traditional sense of an illness that is evidenced by behavioral problems, or mental illness as determined by symptoms, or mental illness as determined by the precipitating cause. The policy’s definition is not helpful. It merely begs the question by essentially defining mental illness as a mental disease or disorder.
Based upon these considerations, we hold that the term “mental illness” and its definition are ambiguous in that they are susceptible to more than one reasonable interpretation. See Keller v. Safeco Ins. Co., 317 Ark. 308, 877 S.W.2d 90 (1994). We therefore conclude that summary judgment in favor of First Unum was improper in this case.
Ordinarily, if there are varying interpretations to be accorded a provision in an insurance policy, one favoring the insurer and another favoring the insured, the one favoring the insured will be adopted. See id. However, we decline to adopt, as a matter of law, appellant’s argupient that the term “mental illness” must be defined by reference to the cause of the illness. First, the term is not susceptible to any one interpretation, as shown by our previous discussion. Secondly, the ‘.‘plain, ordinary and popular” meaning of the term and its definition cannot be determined without reference to extrinsic evidence, making it improper for interpretation as a matter of law. See Smith v. Prudential Prop. & Cas. Co., supra. Finally, the approach used to interpret the policy must be fluid enough to account for the different thoughts among the medical and lay communities regarding the nature of mental illness and to account for new discoveries and changes in attitude that accompany modern medicine. Considerations such as these necessarily involve a factual finding, and that is what we, determine is required *62in this case. We therefore reverse and remand to allow a fact-finder to determine whether Elam’s disorder is or is not a mental illness.
We realize that our approach may require a determination on virtually a case-by-case basis of whether a particular disorder is a mental illness. However, in the absence of a more specific policy definition, and considering that no one factor may be conclusive in dealing with this question, this is the most' satisfactory solution to a very difficult problem.
In light of our remand, we do not reach the question of whether appellant’s Hepatitis B and migraine headaches have any bearing on his disability.
Reversed and remanded.
Koonce and Meads, JJ., agree.