This is an appeal by the plaintiff, Della Faye Gilbert, from a summary judgment denying her relief in an action for breach of an employment contract against the defendant, Fireside Enterprises, Inc. Fireside’s motion was based solely on the ground that appellant’s action was precluded by virtue of a prior judgment in a negligence action between the same parties arising out of the same transaction.1 Thus, the trial judge could only have rendered summary judgment on this ground. We reverse on the ground that appellant’s claim is not precluded by the prior judgment on her suit for damages based upon negligence because the present cause of action, although arising out of the same occurrence is based upon contract.
Appellant brought the present suit on May 10, 1978, alleging that Fireside was a subscriber to the Texas Workers’ Compensation Act at the time she was hired by Fireside; that Fireside subsequently breached an implied term of its employment contract with her by failing to maintain its subscription to the act without giving notice as prescribed in Tex.Rev.Civ.Stat.Ann. art. 8306, § 3c (Vernon 1967); and that Fireside is liable, therefore, to her for an amount equivalent to the workers’ compensation benefits she would have received under the act, had the subscription not expired, for personal injuries sustained by her while she was engaged in her duties as Fireside’s employee.2 In response, Fireside moved for summary judgment on the ground that a prior judgment in a suit between the parties arising out of the same occurrence precludes appellant from asserting her claim in the present action. In that suit, filed by appellant on May 1, 1975, appellant sought damages against Fireside for her injuries, which she alleged were sustained as a result of Fireside’s negligence. Following an adverse jury verdict, judgment was rendered against her. She appealed that judgment to this court and we affirmed.3
Appellant argues that the doctrine of res judicata does not preclude her present suit because, although arising out of the same occurrence, and although between the same parties, the present suit is one for breach of contract, whereas the prior suit was a tort claim. On the other hand, Fireside contends that appellant’s cause of action should be precluded under the doctrine of res judi-cata, because with due diligence she could have litigated her present action for breach of contract in the prior negligence action. Thus, the question presented is whether appellant’s present cause of action in contract should be barred under the doctrine of res judicata by the final judgment rendered against her in the prior tort action. We *871conclude that res judicata does not preclude her present action in contract.
Under the doctrine of res judicata, a valid and final personal judgment on the merits of a plaintiff’s cause of action precludes the plaintiff from thereafter maintaining an action on that cause of action. Restatement of Judgments, §§ 47,48 (1942). This is a restatement of the traditional res judicata principles of bar (applicable when the judgment is rendered in favor of the defendant) and merger (applicable when the judgment is rendered in favor of the plaintiff), which are rules of claim preclusion. Steakley & Howell, Ruminations on Res Judicata, 28 SW.L.J. 355, 356 (1974). Bar and merger, or “technical” res judicata, are to be distinguished from collateral es-toppel, which is a rule of issue preclusion, but the two are often confused. Id. at 356-57; see Franklin v. Rainey, 556 S.W.2d 583, 585 (Tex.Civ.App.—Dallas 1977, no writ). For purposes of clarity, we shall use the term “res judicata” to refer only to claim preclusion.
The difficulty encountered in applying the doctrine of res judicata turns on how a court defines the boundaries of a particular “cause of action.” This is true because all courts generally agree that a prior judgment precludes relitigation of all issues actually considered in the prior suit as well as all issues which should have been considered between the same parties or their privies.4 E. g., Texas Water Rights Commission v. Crow Iron Works, 582 S.W.2d 768, 771-72 (Tex.1979); Griffin v. Holiday Inns of America, 496 S.W.2d 535, 538 (Tex.1973); Abbott Laboratories v. Gravis, 470 S.W.2d 639, 642 (Tex.1971); Ogletree v. Crates, 363 S.W.2d 431, 435 (Tex.1963). Thus, the problem is determining how many issues which could have been litigated in the prior suit, but were not, have been merged into, or barred by, the prior judgment. See Griffin v. Holiday Inns of America, 496 S.W.2d 535, 538 (Tex.1973).
In this latter respect, the problem is in determining which of those issues that might have been, but were not actually, adjudicated in the prior action are precluded from being litigated in the subsequent action. Griffin v. Holiday Inns of America, 496 S.W.2d at 538. This is precisely the problem that confronts us in the instant case, since the relevant issues in appellant’s present claim might have been, but were not actually, adjudicated in her prior negligence action against Fireside.
In the frequently-cited case of Freeman v. McAninch, 87 Tex. 132, 27 S.W. 97 (1894), the supreme court, quoting Ohio authority, stated that, under the doctrine of res judi-cata: “A party can not relitigate matters [i. e., issues] which he might have interposed, but failed to do, in a prior action between the same parties or their privies, in reference to the same subject matter.” 87 Tex. at 139, 27 S.W. at 100. Accord, Abbott Laboratories v. Gravis, 470 S.W.2d 639, 642 (Tex.1971). Elsewhere in the Freeman opinion the court expressly reaffirmed the propositions enunciated in the leading case of Foster v. Wells, 4 Tex. 101 (1849), in which the court expressed the doctrine of res judicata in the following terms:
The general proposition that the judgment or decree of a court possessing competent jurisdiction shall be final as to the matters determined, can not be controverted. “The principle, however, extends further; it [the judgment] is not only final as to the matters actually determined, but as to any other matter which the parties might litigate in the cause, and which they might have had decided.” But it is only when the trial was on the merits, where all the matters between the parties were or could have been adjudicated, that the judgment is a bar [or *872operates as a merger] to another action. [Emphasis added.]
4 Tex. at 104 (citations omitted).
Read literally, this language appears to suggest that the doctrine of res judicata precludes litigation in a subsequent action of all issues, which, with the use of diligence might have been, but were not actually, adjudicated in a prior action between the same parties or their privies. Under more recent Texas cases, however, the doctrine is not so broad. Rather, the doctrine of res judicata under current Texas case law precludes litigation in a subsequent action of issues which, with the use of diligence, might have been, but were not actually, adjudicated in a prior action between the same parties or their privies if, and only if, such issues constituted part of, or a defense to, a cause of action validly and finally adjudicated on the merits in the prior action.
As the supreme court noted in Griffin v. Holiday Inns of America, 496 S.W.2d 535 (Tex.1973): “Freeman has been consistently cited for the proposition that all grounds of recovery or defense relating to the cause of action asserted in the pending suit must be urged or will be barred by the judgment.” 496 S.W.2d at 537-38. In support of this statement, the court pointed to Ogletree v. Crates, 363 S.W.2d 431 (Tex.1963), which, along with Abbott Laboratories v. Gravis, 470 S.W.2d 639 (Tex.1971), cited Freeman for the proposition that: “The rule of res judicata in Texas bars litigation of all issues connected with a cause of action or defense which [issues], with the use of diligence, might have been tried in a former action as well as those which were actually tried.” 363 S.W.2d at 435; 470 S.W.2d at 642. In further support of its statement, the court in Griffin additionally cited Moore v. Snowball, 98 Tex. 16, 81 S.W. 5 (1904), in which the supreme court interpreted the language of Freeman as meaning “only that all matters which properly belong to a cause of action asserted in the pending suit such as will sustain or defeat, in whole or in part, that cause of action, must be produced or be barred by [or merged into] the judgment.” 98 Tex. at 24, 81 S.W. at 8. Again citing Moore, the court in Griffin noted conversely that: “[A] judgment on the merits in a suit on one cause of action is not [under the doctrine of res judicata] conclusive of [any issue in] a subsequent suit on a different cause of action except [under the principle of collateral estoppel] as to issues of fact actually litigated and determined in the first suit.” 496 S.W.2d at 538. Referring to Ogletree as an exception to this last proposition, the court in Griffin characterized the above statement as a general rule. Id. But the supreme court in its earlier decision in Ogletree expressly stated that the broad causes of action in the prior and subsequent actions under consideration were the same. 363 S.W.2d at 436.
Superficially, it might appear that the broader view of the scope of the doctrine of res judicata, as expressed in Ogletree, was reaffirmed by the supreme court in Texas Water Rights Commission v. Crow Iron Works, 582 S.W.2d 768 (Tex.1979), in which the court, relying on Griffin, Gravis and Ogletree, stated:
The scope of res judicata is not limited to matters actually litigated; the judgment in the first suit precludes a second action by the parties and their privies not only on matters actually litigated, but also on causes of action or defenses which arise out of the same subject matter and which might have been litigated in the first suit. [Emphasis added.]
582 S.W.2d at 771-72. Elsewhere in the Crow Iron Works opinion, however, the court noted: “The doctrine of res judicata states that a cause of action once finally determined, .. . between the parties, on the merits, by a competent tribunal, cannot af-terwards be litigated by new proceedings either before the same or any other tribunal.” Id. at 771. In view of this language and the court’s reliance on Griffin, Gravis and Ogletree as authority for its holding, it appears that the supreme court in Crow Iron Works did not intend to indorse a broader view of the scope of the doctrine of res judicata than that previously expressed in Griffin.
*873Thus, in summary, Texas law conforms to the observations of Professor Cleary, who noted:
A literal reading of the rule that res judicata applies not only to what was litigated but to what might have been litigated, as well, would mean that all procedurally joinable matters between parties at the time of the former action would now be res judicata, regardless of how unrelated such matters might be in fact. Courts have not gone to that length. They have said that what might have been litigated in the first action is res judicata only to the extent that it constituted a part of the cause of action involved in the first action. [Emphasis added.]
Cleary, Res Judicata Reexamined, 57 Yale L.J. 339, 346 (1948). As the court in Griffin clearly recognized, the problem in this respect is in determining whether the subsequent and prior actions under consideration concern the same cause of action or different causes of action. 496 S.W.2d at 538. The precise question before us is, therefore, whether the issues in appellant’s present action — which, with the use of diligence, might have been, but were not actually, adjudicated in her prior negligence action— constituted part of the cause of action asserted by her in her former action such that they are barred from being presently litigated under the doctrine of res judicata.
Thus, the question of whether a prior judgment is res judicata turns on the definition of a cause of action. Leading commentators have enunciated several different concepts of what constitutes a cause of action for purposes of determining the applicability of the doctrine of res judicata in a particular case. See generally Schopflocher, What is a Single Cause of Action for the Purpose of the Doctrine of Res Judicata?, 21 Ore.L.Rev. 319, 320-23 (1942); Steakley & Howell, Ruminations on Res Judicata, 28 SW.L.J. 355, 360-63 (1974). The narrowest of these is the individualized concept, under which a cause of action consists of a “material set of facts which is sufficient to individualize, under a specific rule of the substantive law, plaintiff’s claim so as to distinguish this claim from any other claim against the same defendant.” Schopflocher, supra at 322. Expressed differently, under this concept a cause of action consists of a “group of operative facts which, standing alone, would show a single right in the plaintiff and a single delict to that right giving cause for the state, through its courts, to afford relief to the party or parties whose right was invaded.” McCaskill, Actions and Causes of Action, 34 Yale L.J. 614, 638 (1925). Thus, under this concept an additional cause of action exists only when there exists one or more additional material facts which make available to the plaintiff recovery under a different specific rule of substantive law. Steakley & Howell, supra at 361.
A broader concept of what constitutes a cause of action for purposes of applying the doctrine of res judicata is the pragmatic concept, under which a cause of action consists of a “set of facts which, according to human experience and usage and trial convenience, may fairly be deemed a unity, and which, if true, under any rule of the substantive law whatsoever, justifies plaintiff’s claim for relief.” Schopflocher, supra at 323. Under this concept, the set of facts constituting a cause of action “will give rise to at least one right of action, but it is not limited to a single right (if it is ever possible to isolate one such right from others).” C. Clark, Code Pleading 137 (2d ed. 1947). Expressed differently, under this concept an additional cause of action does not exist, even though there exists one or more additional facts which make available to the plaintiff recovery under a different specific rule of substantive law, if the additional facts are so closely related to the facts which otherwise constitute an individualized cause of action so as to justify consideration of all the facts as one operative unit. Schopflocher, supra at 322-23. In essence, under this concept “[t]he extent of the cause [of action] is to be determined pragmatically by the court, having in mind *874the facts and circumstances of the particular ease.” C. Clark, supra at 137.5
The broadest concept of what constitutes a cause of action for purposes of determining applicability of the doctrine of res judi-cata is the procedural duty concept.6 This concept contemplates that a party should be required to join all closely-related claims in one action. Steakley & Howell, supra at 363. A cause of action under this concept consists of all claims arising out of the same transaction or occurrence, essentially similar to that under which a compulsory counterclaim is defined. See Tex.R.Civ.P. 97(a).
The supreme court has expressly rejected the procedural duty concept of what constitutes a cause of action for purposes of determining applicability of the doctrine of res judicata. In Griffin, that court stated:
We have not said or held that a judgment in a suit on one claim or cause of action is necessarily conclusive of all claims and causes of action against the same party, or relating to the same property, or arising out of the same transaction. Apparently some courts have gone that far, but we are unwilling to do so.
496 S.W.2d at 538. Furthermore, careful analysis of the supreme court’s most recent res judicata opinions reveals that the court has not adopted either the pragmatic or individualized concepts. Instead, the supreme court has applied the doctrine of res judicata on a functional ad hoc basis.
In Ogletree, the court held that the doctrine of res judicata barred respondent’s efforts to set aside, on grounds of fraud and duress, a 1957 Alabama judgment awarding petitioner child custody. In 1958 both parties had unsuccessfully attempted to have the judgment modified by the Alabama court. Further, in 1960 respondent had unsuccessfully prosecuted an action in Texas to change custody on grounds of changed circumstances. Finally, in 1961 she brought the action to set aside the 1957 judgment on grounds of fraud and duress. The supreme court noted, however, that her 1961 action was based on facts fully known to her at the time of the 1958 and 1960 proceedings, in which she had affirmed the validity of the 1957 judgment and sought its modification. On this basis, the supreme court held:
There may be a technical distinction between [the individualized cause of action in] a suit to obtain custody and possession of a minor child through modification of a final judgment and [the individualized cause of action in] a suit to obtain custody and possession of the child by setting aside a final judgment, but the broad cause of action and the relief sought in both suits are the same. The suit in each instance tests the rights of the parties to custody of the child and the only pertinent inquiry is the best interests of the child. As a matter of public policy there should be a high degree of stability in the home and surroundings of a young child, and, in the absence of materially changed conditions, the disturbing influence of constant re-litigation should be discouraged. Once a final judgment of custody is rendered, a subsequent suit to modify or to avoid the judgment should be res judicata of all [individualized] causes of action which, with diligence, could have been asserted in the suit as a basis for obtaining custody and possession of the child.
363 S.W.2d at 436.
In Gravis the supreme court held that plaintiff’s subsequent products liability ac*875tion against the defendant was barred by an adverse judgment in plaintiff’s prior negligence action against the same defendant. In this respect the court, citing Ogle-tree noted:
The present action against Abbott [the petitioner] can be construed as one founded wholly upon the theory of products liability. The contention is that the former suit was one based on negligence, and that a final judgment for a defendant in a negligence action would not, even between the same parties, be a bar under the doctrine of res judicata to a recovery in a second suit based on products liability under Section 402A of the Restatement of Torts (2d), and McKisson v. Sales Affiliates, Inc., 416 S.W.2d 787 (Tex.1967). We disagree. Both suits involve a tort action resulting from the furnishing of the same drug for the same operation on the same person.
The claim for recovery against Abbott in the first action could [with the use of diligence] have been founded upon products liability or negligence, or both. The final judgment for the defendant Abbott in the first suit is therefore res judicata in the present suit. [Emphasis added.]
470 S.W.2d at 642.
In Griffin the supreme court expressly reaffirmed the holding in Gravis, stating: “It is generally recognized that a judgment in a suit to recover damages on one theory is conclusive as to all theories of liability that might have been alleged and all defenses that might have been urged with respect to the claim.” 496 S.W.2d at 537. In Griffin, the supreme court held that petitioner’s subsequent quantum meruit action was barred by the judgment in respondent’s prior cross-action against petitioner for breach of contract because the claim in quantum meruit was a compulsory counterclaim to that prior cross-action. Id. at 536. The court further noted, however, that had the claim in quantum meruit not been a compulsory counterclaim in the prior action, it would not have been subsequently barred by virtue of the judgment in petitioner’s own action against respondent for breach of the contract with respect to which the quantum meruit claim was related. In this respect, the supreme court noted:
[A] majority of the Court has concluded that we should adhere to the established rule in this and most other jurisdictions that a judgment for the defendant in a suit for breach of contract on the ground that he is not liable for breach of contract does not preclude a subsequent suit in quantum meruit, the causes of action being regarded as different for res judicata purposes. [Emphasis added.]
Id. at 538.
In Crow Iron Works the supreme court held that the doctrine of res judicata, coupled with the doctrine of lis pendens, precluded petitioners’ efforts to set aside the Texas Water Rights Commission’s order denying their petition to upgrade water rights awarded to their predecessors in interest under a judgment in a prior action lasting approximately fourteen (14) years in which the trial court exercised jurisdiction over water rights to the Rio Grande River in four counties, affecting 850,000 acres of land, and involving approximately 3,000 named defendants. 582 S.W.2d at 771. In so holding, the court, citing Griffin, Abbott and Ogletree, noted:
All the requirements of res judicata have been shown in this cause. Although the record does not reveal whether Crow Iron Works, et al’s predecessors at the time of the inception of the litigation asserted the water rights in question here, they might have asserted these rights. We therefore hold that the judgment in the Valley Water Case is res judicata of the water rights asserted by Crow Iron Works, et al. Most of the policy considerations underlying the doctrine of res judi-cata, including the desirability of stable decisions and economy of court time, require this holding.
Id. at 772.
From the foregoing analysis of supreme court opinions, it is impossible to reconcile *876these cases under the view that they all adhere to the pragmatic concept of a cause of action. Griffin —with its holding that a claim for breach of contract and a claim for quantum meruit are different causes of action even though they arise out of the same contractual transaction — belies this view. Similarly, it is also impossible to reconcile these cases under the view that they all adhere to the individualized concept of a cause of action. Gravis — with its holding to the effect that a claim based on negligence and a claim based on products liability are part of but one cause of action when they arise out of the same tortious transaction — particularly belies this view. Further, it cannot be said that any of these cases supplants, or is controlling over, the others, since each succeeding case has expressly or impliedly reaffirmed the prior decisions. These cases can be reconciled, however, on the basis that the supreme court has been employing a functional concept, rather than rigidly adhering to a mechanical definition of a cause of action under the doctrine of res judicata.
Though not expressly articulated in each of these cases, it is apparent that the supreme court’s functional concept of a cause of action is based on a case by case analysis of relevant policy considerations. Under this functional concept, issues which might have been, but were not actually, adjudicated in a prior action constitute part of the cause of action adjudicated in the prior action if litigation of these issues in a subsequent action between the same parties or their privies is offensive to the policy considerations underlying the doctrine of res judicata. It necessarily follows that under this functional concept, according to which results depend largely on the specific form of the transaction with which the subsequent and prior actions are concerned, the precedential scope of existing res judicata opinions is narrow.
For example, Westinghouse Credit Corp. v. Kownslar, 496 S.W.2d 531 (Tex.1973), confirms that the supreme court is indeed employing this functional concept of a cause of action for the purpose of applying the doctrine of res judicata. In Kownslar, Westinghouse Credit was the obligee on nine promissory notes with respect to which Kownslar had executed a single agreement as guarantor. By March 1967, all nine notes were in default. In May 1967, Westinghouse Credit instituted suit against Kownslar to collect the amount due on four of the notes and obtained a judgment in its favor which was affirmed on appeal. In 1971 Westinghouse Credit sued Kownslar to collect the amount due on the remaining five notes. In response, Kownslar asserted that this subsequent action was precluded under the doctrine of res judicata. In holding that it was not, the supreme court set forth its analysis as follows:
We proceed by two inquiries. We look first to see whether or not there is a Texas case directly in point. We conclude there is not. By application of the narrow construction of [any] precedent [which would tend to preclude the subsequent litigation] appropriate to this doctrine, we find that the case is one of first impression.
Our second problem is whether the factual. situation presented is such that the purposes of the doctrine of merger [or bar] shall be frustrated absent enforcement of the bar. ... Of course, courts are not prepared to hear proof on those subjects and to have the applicability of merger depend upon findings of fact in a particular case. Merger [or bar] follows, if at all, as a consequence of the form of the transaction and our holding herein must apply to all transactions which take this form ....
The form of this transaction is not exactly within the rule of any decided case in our jurisdiction. It does not clearly appear to the court that separate suits would be, more often than not, the engine of oppression and waste. We hold that merger does not bar the present suit.
496 S.W.2d 532-33. To the same effect was the Crow Iron Works opinion, in which the *877court expressly stated that its holding was dictated by “most of the policy considerations underlying the doctrine of res judica-ta.” 582 S.W.2d at 772.
The functional definition of a cause of action under the doctrine of res judicata which the supreme court has adopted perhaps offers a sound approach to problem solution in this area of the law. As Professor Cleary has noted in suggesting that courts adopt such an approach: “Deciding cases by [mere mechanical] definition is easy and therefore a method which courts are prone to follow, but the easy way is not necessarily the best way.” Cleary, supra at 341.
Applying the functional approach adopted by the supreme court, we note that there is no precedent directly on point with the present case. We must proceed to determine whether the prosecution of appellant’s present action is offensive to the policy considerations underlying the doctrine of res judicata and is, therefore, precluded under the doctrine.
The traditional policy considerations underlying the doctrine of res judicata are essentially four in number. These are promotion of judicial economy, prevention of vexatious litigation, prevention of double recovery, and promotion of the stability of decisions. See generally Cleary, supra at 344-49. As the Kownslar opinion indicates, compelling policy considerations must be substantially offended by maintenance of a subsequent action before preclusion of the action is warranted through application of the harsh doctrine of res judicata. See 496 S.W.2d at 532.
The first of these policy considerations, promotion of judicial economy, is generally an uncompelling rationale for applying the doctrine of res judicata in a particular case. We note, however, that in exceptional circumstances, such as those involving massive cases which have tied up court dockets for considerable lengths of time, promotion of judicial economy may be a compelling rationale for invoking the doctrine of res judicata as was the case in Crow Iron Works. But the present case does not present such exceptional circumstances. Thus, we agree with Professor Cleary, who stated:
Courts exist for the purpose of trying lawsuits. If the courts are too busy to decide cases fairly and on the merits, something is wrong. Decision solely in terms of the convenience of the court approaches the theory that the individual exists for the state. Maintenance of the judicial system is a very minor portion of the cost of government. If the judges are too few to be able to decide cases fairly and on the merits, the public probably can afford to have more judges. The fact that a party may waive the defense of res judicata, as seems to be the general rule, indicates that saving the judge’s time is more afterthought than reason.
Cleary, supra at 348-49. We decline, therefore, to apply the doctrine of res judicata in the present case because appellant’s bringing two actions against Fireside, instead of one, might in some way be judicially uneconomical.
We now turn to whether the present action should be barred on premise of preventing vexatious litigation. The term “vexatious litigation” appears to be directed primarily at the situation in which a plaintiff attempts to make two actions do the work of one but does not attempt a double recovery or threaten the stability of a decision. Id. at 346. In this respect, it has been noted:
[T]he purpose of liberality in joinder rules is the same as the anti-vexatious-limitation purpose of the rule of res judicata, i. e., to encourage litigants to reduce the numerical volume of lawsuits by bringing more disputed matters into the same action. Yet when plaintiff seeks to make two lawsuits do the work of one, the rule of res judicata applies too harsh a penalty (complete loss of plaintiff’s right of recovery), and permissive joinder too slight *878a penalty (some added inconvenience and expense to plaintiff, which he incurs voluntarily).
Id. at 346-47. In our view, a more appropriate penalty in this type of situation, one which more nearly fits the punishment to the crime, is assessment of costs. Under Tex.R.Civ.P. 141: “The court may, for good cause, to be stated on the record, adjudge the costs otherwise than as provided by law or these rules.” Thus, the trial judge has authority to impose this penalty for vexatious litigation under existing law. It has further been suggested that this penalty should include additional attorney’s fees and other expenses, such as those required to secure expert testimony, incurred by a defendant and attributable to the fact that a plaintiff has brought more than one action. Cleary, supra at 347. While existing law does not permit these penalties, the subject is one certainly worthy of consideration by the legislature or the supreme court instead of invoking the harsh doctrine of res judicata.
In any event, we decline to apply the doctrine of res judicata in the present case because appellant might have less vexatiously prosecuted her two actions as one. Application of the doctrine solely for this reason would be too harsh a penalty under the circumstances; prevention of vexatious litigation, as we have defined the term, is simply not a sufficiently compelling consideration to warrant precluding her cause of action under res judicata. The truly legitimate and compelling policy considerations underlying the doctrine of res judicata are, in most cases at least, prevention of double recovery and promotion of stability of decisions. As Professor Freeman stated:
A judgment is a bar [or operates as a merger] not because a party has done some act which precludes him from asserting a right or title; it is properly a bar [or operates as a merger] on principles of public policy, because the peace and order of society, the structure of our judicial system, and the principles of our government require that a matter once litigated should not be again drawn in question between the same parties or their privies.
2 A. Freeman, Judgments § 626 (5th ed. 1925).
The danger of double recovery exists in a subsequent action only if the plaintiff has recovered previously. Cleary, supra at 344. In the instant case, appellant was the judgment loser in her prior action against Fireside. Thus, there exists in the instant case no danger of double recovery in permitting appellant to prosecute her present contract claim.
As to promoting the stability of decisions, Professor Cleary observed:
Effective operation of courts in the social and economic scheme requires that their decisions have the respect of and be observed by the parties, the general public and the courts themselves. According insufficient weight to prior decisions encourages disrespect and disregard of courts and their decisions and invites litigation. However, giving a prior decision effect beyond its actual scope is just as undesirable as according it to an insufficient effect.
Pronouncements that courts will not pass upon particular matters because of a highly conceptual view of what might have been decided in a prior action serve neither to enhance the stature of courts nor to add strength to the foundations upon which they rest.
Id. at 345-46. The stability of a decision may be threatened when a plaintiff splits up the underlying fact situation which gives rise to litigation and uses the pieces to build more than one lawsuit. This is not so, however, when a plaintiff splits arithmetically (i. e., selects certain elements of damage for the first action and other elements for the second), as was the case in Kowns-lar, or when a plaintiff splits the relief requested. Id. at 346. See also id. at 343-44. On the other hand, stability of a decision is quite arguably threatened when a *879plaintiff splits theories of recovery, as in the case before us. Id. at 346. As Professor Cleary noted:
Policy would seem to indicate that when a plaintiff has once attempted to obtain his entire relief, based upon his entire damages, then the matter should be laid at rest. He should be denied a second attempt at substantially the same objective under a different guise. In deciding such cases, the concept of a cause of action possesses real utility if applied with realization of what actually is at issue ....
Id. at 346.
It would appear that this is the rationale underlying the result in the Gravis case— with the result in Ogletree resting on substantially similar grounds, reinforced, doubtless, by the particularly significant policy considerations peculiar to suits concerning the parent-child relationship. It is also apparent that this is the rationale underlying the supreme court’s reaffirmance of Gravis and Griffin with the statement that: “It is generally recognized that a judgment in a suit to recover damages on one theory is conclusive as to all theories of liability that might have been alleged and all defenses that might have been urged with respect to the claim.” 496 S.W.2d at 537. However, for unexplained reasons of policy, the court in Griffin decided to adhere to precedent which, in effect, dictated a result contrary in principle to that reached in Gravis.
The conflict in principle between Griffin and Gravis essentially concerns a conflict of views as to the proper scope to be accorded the doctrine of res judicata in promoting stability of decisions through preclusion of the splitting of theories of recovery. In this respect, splitting may be as to either factual theories of recovery or legal theories of recovery or both. Cleary, supra at 343. The supreme court in Gravis was apparently of the view that splitting in either fashion threatens the stability of decisions. On the other hand, the supreme court in Griffin was apparently of the view that splitting as to factual theories of recovery threatens the stability of decisions but splitting as to legal theories of recovery does not. In no other way is it possible to explain, in policy terms, why the Gravis court held that a prior negligence action barred a subsequent products liability action arising out the same fact situation, while the Griffin court stated that a plaintiff’s prior contract action did not bar his subsequent action in quantum meruit arising out of the same fact situation.
Under the rule enunciated in Kownslar, preclusion precedent, such as Gravis, is not binding unless it is directly on point. As we have noted, no such precedent directly on point with the instant case exists. Accordingly, we conclude that the stability of the decision in a negligence action is not threatened, in the absence of a danger of double recovery, by the maintenance of a subsequent action between the same parties for contractual recovery of the equivalent of workers’ compensation benefits. Thus, we hold that the policy of promoting stability of decisions does not mandate that appellant's present action be barred under the doctrine of res judicata. In this respect, we note that appellant has split legal, rather than factual, theories of recovery in her two actions against Fireside. She is in no way presently attempting to urge that Fireside is liable to her in negligence under a factual theory different from that urged in her prior action against Fireside. It can hardly be argued, therefore, that the stability of the decision finding Fireside not negligent in that prior action is threatened by appellant’s maintenance of her present action in contract. In our view, the stability of a decision is normally threatened only when the legal theory of recovery in a prior and a subsequent action remains the same, or substantially the same, and different factual theories, stemming from the underlying situation giving rise to the litigation in the first place, are asserted in the prior and subsequent actions to justify recovery under the same legal theory. This is the view *880implicit in Griffin and, as a general rule, the stability of a decision is sufficiently promoted by the narrower scope accorded the doctrine of res judicata in that case. We question the wisdom of giving the doctrine the broader scope accorded it in Gra-vis, although we recognize that this may be warranted in cases such as Ogletree which concern special subject matter and highly sensitive policy considerations, such as child custody.
Since we have concluded that appellant’s maintenance of her present action against Fireside does not substantially transgress upon a compelling policy consideration underlying the doctrine of res judicata, we hold that the doctrine does not bar her present action. Accordingly, we reverse the judgment of the trial court and remand the case for trial.
. The doctrine of election of remedies, although advanced here by Fireside, was not presented to the trial court as required by Tex.R.Civ.P. 166-A. Thus, we do not address that question.
. The cause of action appellant is asserting is recognized in Anderson-Berney Realty Co. v. Soria, 123 Tex. 100, 67 S.W.2d 222 (1933).
. Gilbert v. Fireside Enterprises, Inc., No. 75-4374-A (Dist.Ct. of Dallas County, 14th Judicial Dist. of Texas, August 12, 1977), aff'd No. 19535 (Tex.Civ.App.-Dallas, May 24, 1978, no writ).
. Litigation of such issues is not precluded, however, under the principle of collateral es-toppel; collateral estoppel only precludes reliti-gation of issues actually and finally adjudicated in a prior action. Hanrick v. Gurley, 93 Tex. 458, 480, 56 S.W. 330, 330 (1900), citing Cromwell v. County of Sac, 94 U.S. 351, 352-53, 24 L.Ed. 195 (1876). See generally Steakley & Howell, supra note 3, at 356-57.
. What may be characterized as a somewhat different pragmatic concept of a cause of action is that set forth in J. Pomeroy, Code Remedies: Remedies and Remedial Rights §§ 346-47 (5th ed. 1929). See generally Steakley & Howell, supra note 3, at 362 n.60.
. Under this concept, the preclusion from asserting a second action is based upon a procedural duty to join actions, not upon the doctrine of res judicata, thus permitting a court to decline to preclude a second action in the absence of fault in failing to join such action in a prior action. Steakley & Howell, supra note 3, at 363. However, this distinction is largely one of semantics, since the doctrine of res judicata under Texas law precludes litigation of only those issues which, with the use of diligence, might have been litigated in a prior action. Abbott Laboratories v. Gravis, 470 S.W.2d at 642; Ogletree v. Crates, 363 S.W.2d at 435.