The plaintiff, John Mole, a former tenured professor at the University of Massachusetts Medical Center (UMMC), appeals from the granting of a directed verdict in favor of all defendants on his claims of retaliation (under G. L. c. 151B, § 4 [4] and [4A], and Title VII, 42 U.S.C. § 2000e-3 [2000]) *583and violation of civil rights (42 U.S.C. § 1983 [2000]).2 The Appeals Court, with one Justice dissenting, reversed and remanded the case for a new trial. Mole v. University of Mass., 58 Mass. App. Ct. 29, 48 (2003). We granted the defendants’ application for further appellate review. We conclude that the plaintiff failed to introduce sufficient evidence to establish the requisite causal connection between his protected activity (supporting his wife’s claim of sexual harassment) and the adverse personnel actions of which he complains. We therefore affirm the judgment.
1. Facts. Viewed in the light most favorable to the plaintiff, the evidence at trial was as follows. Professor Mole and his wife, Jacqueline Anderson, were both employed at UMMC as faculty members in the biochemistry and molecular biology department (department). Michael Czech was the chair of the department, and Frank Chlapowski was the acting chair during periods when Czech was on leave of absence. With the full support of Czech, Mole had joined the faculty in 1981, been granted tenure in 1984, and attained the rank of full professor in 1987. Anderson was not in a tenure track position.
Mole and Anderson were the founders and directors of a research laboratory at UMMC known as the Protein Chemistry Core Facility (PCF), which supported the research of other scientists at UMMC (as well as scientists from outside the university) by isolating and sequencing proteins and amino acids. Funding for the PCF was provided by two sources, the Diabetes and Endocrinology Research Center (DERC), a group of research scientists at UMMC that in turn was funded by Federal grants, and the Scientific Council, another group of UMMC scientists comprised of one member from each department at UMMC. In 1988, Czech criticized Mole with respect to certain fees that the PCF was charging, alleging that the fees were “flagrantly excessive.” Czech expressed dismay about the fees in light of the support the PCF received from the Scientific *584Council, and disappointment that the imposition of such fees “reflects a lack of cooperative spirit and collegiality.” In response to this criticism, Mole and Anderson wrote to Czech, opining that it was “clear” that Czech had “embarked on a course of action that is intended to damage us personally and professionally” and accusing Czech of making “misleading and false accusations” against them. They indicated that, unless their concerns were resolved, they would file a formal grievance. Dissatisfied with Czech’s response, they pursued that grievance, seeking to remove from their respective personnel files Czech’s correspondence containing the alleged “false and misleading information.” The grievance was resolved in August, 1988, with Czech agreeing that his letter was a “private communication” that was not intended to be placed in their personnel files.
Beginning in the late 1980’s, the Scientific Council questioned the advisability of its support of the PCF. At the time, various researchers voiced concerns about difficulties they had encountered working with Mole, and some researchers were opting to have samples analyzed by other facilities instead of using the PCF. A subcommittee of the Scientific Council, formed to review alternative proposals to the PCF, rendered its report on February 8, 1990, recommending that the PCF be merged with another nearby facility and that a new director be recruited for the new combined laboratory.3
Three weeks later, on February 28, 1990, Anderson submitted an informal complaint to the university’s equal employment opportunity office, alleging that Czech had made sexually offensive remarks. In some unspecified manner, Mole had supported or encouraged his wife’s decision to make that complaint. It is unclear how soon thereafter Czech or Chlapowski learned of Anderson’s complaint, or when they learned of Mole’s own involvement in supporting his wife’s pursuit of that complaint, but, by sometime in August, 1990, both Czech and Chlapowski were aware of the pending complaint.
Meanwhile, back in the late spring of 1990, Mole’s teaching assignments had been curtailed, after receiving negative com*585ments from student evaluations.4 Some negative comments concerning Mole’s teaching had emerged in student evaluations in prior years, but Czech had until that time discounted those negative evaluations for various reasons.
On October 4, 1990, Czech and Chlapowski gave Mole his evaluation for the academic year July, 1989, through June, 1990. Whereas the prior year’s evaluation had praised his teaching as “very helpful to the department” and noted improvement in his lectures, the 1990 evaluation referenced the negative evaluations from students and reflected the fact that Mole had been relieved of his teaching duties. With regard to Mole’s research, the prior year’s evaluation had noted that one of Mole’s outside grants had ended, and that he was in his final year of another grant. That 1989 evaluation had urged Mole “to go forward with strong efforts to ensure continuation of his research funding.” The 1990 evaluation noted that the one remaining grant had not been renewed, but that Mole had pending one grant application to two other organizations. The evaluation referenced the urgency of Mole’s finding new funding for the PCF, because the Scientific Council’s withdrawal of support meant that the PCF had lost nearly one-half of its customary funding. The evaluation expressed disappointment that the Scientific Council had “lost confidence” in Mole’s direction of the PCF, and advised Mole that his “primary focus and concern” for the coming year should be to seek out new sources of research funding.
In January, 1991, Anderson lodged a formal complaint of sexual harassment against Czech. Chlapowski was under the impression (ultimately shown to be mistaken) that he had also been named in the complaint. On April 24, 1991, Chlapowski filed a charge of “scientific misconduct” against Mole based on Mole’s inaccurate listing of certain papers as “accepted for publication” or “in press.” Amongst research scientists, a charge of “scientific misconduct” is extremely serious, and potentially ruinous to one’s career. After investigation by UMMC officials, *586it was determined that Mole’s misreporting of the status of his publications was “sloppy and inappropriate,” and deserving of reprimand, but that it did not amount to “scientific misconduct.” Mole was reprimanded but, in the opinion of one of the investigators, it had been “ridiculous” for Chlapowski to escalate the charge to one of “scientific misconduct.”
In May, 1991, DERC voted to cease its funding of the PCF as of November, 1991.5 Chlapowski then contacted the Scientific Council, asking that it finalize its position with respect to its own funding of the PCF.6 The Scientific Council formally voted to terminate its financial support of the PCF. Shortly thereafter, Mole submitted a request that the department fund the PCF. Chlapowski responded with a list of questions concerning Mole’s request, while making clear that there was no commitment to provide department funding. The questions included requests for factual data (e.g., information on the PCF’s past finances), but also probed the merits of whether support from the department was warranted (e.g., asking why the department should fund the PCF when support “has been discontinued by two of the major scientific programs of the Medical School”). Mole, not wishing to expend the considerable time it would take to gather the information necessary to respond and doubtful that his request for support would be granted, did not supply the requested information.
In December, 1991, Czech and Chlapowski completed Mole’s annual evaluation for the 1990-1991 academic year. That evaluation reflected the fact that Mole had lost all outside sources of funding, and that his funding from DERC and the Scientific Council was at an end. It also recited the history leading to the reprimand for the inaccurate reporting of the publication status of his research, and noted that Mole had no new papers either published or accepted for publication. The evaluation went on to list complaints from postdoctoral students who had worked in Mole’s laboratory. In that evaluation, Czech and Chlapowski *587expressed concern over the fact that Mole had refused to meet with Chlapowski to discuss his evaluation, necessitating resort to purely written communications, and that Mole had similarly alienated other colleagues. In light of these problems, the evaluation concluded that Mole’s “behavior and poor performance are totally inconsistent with his appointment as a tenured Professor,” and that it was “difficult for [Czech and Chlapowski] to understand how Dr. Mole can hope to continue as a faculty member.” In light of these developments, they advised Mole that the laboratory space for the former PCF and Mole’s own laboratory space might be devoted to other uses.
On February 10, 1993, Chlapowski completed the next annual evaluation of Mole, and the jury could infer that Czech (despite being on a leave of absence) had been consulted with respect to that evaluation and concurred in it. The evaluation again reflected that Mole had lost all sources of outside funding, and noted that prior evaluations had advised him to concentrate his efforts on obtaining such funding. Chlapowski recounted that Mole had proposed to submit a grant application that included Anderson as one of the principal investigators, but that he (Chlapowski) had declined to approve it because Anderson would not be a member of the faculty during the time period of the grant.7 Mole had not prepared any other grant applications since May, 1991. The evaluation also criticized Mole for not following through with respect to the correction of his curriculum vitae (a recommendation that had stemmed from the prior accusation of “scientific misconduct”), and for not coming forward with any new publications or research papers. The evaluation cited Mole’s absence from weekly department seminars, indicative of “a further lack of interest on his part in developing a credible research program.” Ongoing problems with his teaching performance were noted, with the conclusion that it would be “inappropriate” to increase his teaching load. The evaluation concluded that Mole had failed to address any *588of the concerns raised in the prior year’s evaluation, and that his “overall productivity” remained “consistently poor.” As a result, Chlapowski advised that Mole’s assigned laboratory space would be consolidated into a single laboratory.8 Finally, Chlapowski announced his intention to recommend to the dean that, unless Mole demonstrated improvement in the coming year, Mole’s salary should be reduced by 17.5 per cent.
On May 2, 1993, Mole filed a charge with the Massachusetts Commission Against Discrimination complaining of alleged retaliation on account of his support for his wife’s charge of sexual harassment.
With respect to the negative performance evaluation he had received in February, 1993, Mole also filed a grievance, claiming that Chlapowski was biased against him and could not evaluate him fairly. The grievance committee recommended that, instead of Chlapowski performing Mole’s evaluation, a committee should be established to perform that task, with the department electing one member of the committee, Mole selecting a second member, and those two members selecting the third member. Mole, of the view that the entire department was biased against him, refused to select anyone from his department to serve on the committee. Lacking any nominee from Mole, the department, at a faculty meeting, voted for two members of the committee, and those two members selected the third member. Neither Czech nor Chlapowski was on the newly formed evaluation committee, nor did Mole present any evidence that Czech or Chlapowski had influenced the faculty vote.
On November 8, 1993, nine faculty members (comprising approximately one-half of the department) signed a letter to the chancellor of UMMC complaining about Mole’s lack of productivity, and protesting that he was occupying laboratory space and receiving a salary “without making any contribution” to the department. The signatories to the letter noted that Mole’s research remained without funding “because, for some *589years now, he has refused to write grant applications. In short, Dr. Mole is a deliberately unproductive faculty member who should be asked to leave.”9 The letter opined that there were grounds for Mole’s dismissal, suggested that Mole be requested to leave, and, if he did not leave voluntarily, “that a formal investigation into his behavior be initiated.” Neither Czech nor Chlapowski signed this letter, and Mole adduced no evidence suggesting that either Czech or Chlapowski had encouraged faculty members to author or sign the letter.
On June 9, 1994, the evaluation committee elected by the department faculty rendered its first evaluation of Mole. Mole, while protesting that the committee was biased against him, submitted to the three evaluators his version of events, explaining why, in his view, he had been unfairly treated by Chlapowski and unfairly denied opportunities to teach and to serve on committees. The evaluators noted that there were “irreconcilable differences of opinion” between Chlapowski and Mole, but concluded that there were grounds for reducing Mole’s teaching and declining to appoint him to committees. Specifically, the evaluation committee noted that there had been “repeated poor performance evaluations” by students in one course, and that Mole had become “a source of general dissatisfaction with the course by students.” With regard to committee assignments, the evaluators recited Mole’s “severely strained relations” with the department faculty, such that “a majority of [department] faculty members would not be willing to be represented by Dr. Mole on any committee.” The evaluation reflected that Mole had no “significant research funding” since 1991, that he had misrepresented the current status of one grant application,10 an infraction that they viewed as comparable to the prior misstatements about the status of his publications. The committee expressed the view that Mole was “not making serious attempts *590at regaining significant research support.” Concluding that Mole was “not an effective faculty member,” the committee recommended that Mole’s salary be decreased by 17.5 per cent, with a further reduction the following year if he did not present “significant evidence of research productivity.” They also recommended that, if Mole did not secure any research funding in the coming year, his laboratory space should be “eliminated” effective July 1, 1995.
When the same committee evaluated Mole again in June, 1995, the evaluators found no improvement. Mole’s only new publication was a paper that had been completed the previous year. Although Mole had submitted two “small” grant applications, neither had been funded, and the committee was of the view that Mole had deliberately sabotaged one of the applications.11 The committee opined that Mole “has not presented any evidence of significant research productivity or made concerted or serious attempts to obtain funding.” Accordingly, consistent with their recommendations in the previous evaluation, they recommended that Mole’s salary should be reduced by another 17.5 per cent and that his laboratory space be assigned to other uses.
No evaluation of Mole was performed the following year. The committee resigned, expressing the view that there was no point in evaluating the work of a faculty member who was doing no work. Mole’s own self-evaluation reflected that he had no new grants or publications, but attached his explanation that the department was denying him “any opportunities to teach, perform service or maintain a credible research program.”
In 1997, the chancellor of UMMC commenced proceedings to terminate Mole.12 At the termination hearings, both Czech and Chlapowski testified and, at the chancellor’s request, Czech *591reviewed and responded to testimony in the hearing transcripts. The chancellor recommended to the board of trustees that Mole be terminated, and the board adopted his recommendation. Mole’s appointment was terminated as of May 5, 1999.
2. Discussion, a. Retaliation. Mole contends that UMMC, Czech, and Chlapowski took the above recited actions against him in retaliation for his support of Anderson’s charge of sexual harassment against Czech. See G. L. c. 151B, § 4(4) and (4A); Title VII, 42 U.S.C. § 2000e-3. Lacking any direct evidence of a retaliatory motive, Mole had the burden of establishing a prima facie case of retaliation, and, in the wake of the defendants’ introduction of nonretaliatory reasons for the various actions taken, the burden of proving that the articulated nonretaliatory reasons were pretext. See McMillan v. Massachusetts Soc’y for the Prevention of Cruelty to Animals, 140 F.3d 288, 309 (1st Cir. 1998), cert. denied, 525 U.S. 1104 (1999); Mesnick v. General Elec. Co., 950 F.2d 816, 827 (1st Cir. 1991), cert. denied, 504 U.S. 985 (1992).
To make out his prima facie case, Mole had to show that he engaged in protected conduct,13 that he suffered some adverse *592action,14 and that “a causal connection existed between the protected conduct and the adverse action.” Mesnick v. General Elec. Co., supra. The defendants contended in their motions for directed verdict, and the judge agreed, that Mole had not introduced sufficient evidence to establish a causal link between his support of his wife’s charge and the actions taken against him, and thus had insufficient evidence to permit an inference of retaliation.
Mole contends that the causal relationship between his support of Anderson’s complaint and the adverse employment actions taken can be inferred from the timing and sequence of events. Of course, “[t]he mere fact that one event followed another is not sufficient to make out a causal link.” MacCormack v. Boston Edison Co., 423 Mass. 652, 662 n.11 (1996), citing Prader v. Leading Edge Prods., Inc., 39 Mass. App. Ct. 616, 617 (1996). That an employer knows of a discrimination claim and thereafter takes some adverse action against the complaining employee does not, by itself, establish causation. “Were the rule otherwise, then a disgruntled employee, no matter how poor his performance or how contemptuous his attitude toward his supervisors, could effectively inhibit a well-deserved discharge by merely filing, or threatening to file, a discrimination complaint.” Mesnick v. General Elec. Co., supra at 828. However, if adverse action is taken against a satisfactorily performing employee in the immediate aftermath of the employer’s becoming aware of the employee’s protected activity, an inference of causation is permissible. See id. (“temporal proximity of an employee’s protected activity to an employer’s adverse action” is circumstantial evidence that allows retaliation claim to get past summary judgment); Oliver v. Digital Equip. Corp., 846 F.2d 103, 110 (1st Cir. 1988) (discharge occurring *593“soon after” employee engages in protected conduct “is indirect proof of a causal connection between the firing and the activity because it is strongly suggestive of retaliation”).
However, the sequence and timing of events in the present case do not support such an inference. With regard to the sequence, Mole’s problems at UMMC began prior to Anderson’s lodging of any charge, and other problems emerged prior to the time that either Czech or Chlapowski knew of that charge or Mole’s involvement in it. Acrimony between Mole and Czech is evident as early as 1988, two years prior to Anderson’s informal complaint of sexual harassment. Although ultimately resolved during the course of Mole’s grievance, the correspondence surrounding that 1988 grievance belies Mole’s theory that all was well in his relationship with Czech prior to the lodging of Anderson’s complaint.
Lack of financial support for Mole’s research — which ultimately became one of the recurring issues referenced in later evaluations and a key factor cited in support of Mole’s salary reductions and termination — also emerges as a problem prior to Anderson’s filing of any charge. Prior to that date, the Scientific Council, which provided nearly one-half of the funding for the PCF, expressed its dissatisfaction with Mole, appointed a subcommittee to review its relationship with the PCF, and had already received the subcommittee’s recommendation that the PCF be merged with another entity and that a new director be found.
The curtailment of Mole’s teaching activities, another event that Mole labels as retaliatory, first occurred in the spring of 1990. While Anderson had by then lodged her informal complaint, Mole has no evidence to show that either Czech or Chlapowski knew about the informal complaint until sometime in the late summer of 1990. See Lewis v. Gillette Co., 22 F.3d 22, 25 (1st Cir. 1994) (absent evidence that other employees knew of plaintiff’s testimony at arbitration hearing, evidence that employees began gawking and staring at plaintiff shortly thereafter did not establish retaliation). Evidence that, at some unspecified point in time, general “rumors” began to circulate about a “female faculty member” complaining of sexual harassment does not suffice to fix an earlier date for the defendants’ *594initial awareness of Mole’s protected activity. Speculation that the defendants might have heard these “rumors,” and that they might have deduced who the complainant was, who she was complaining about, and what role Mole had played in that complaint, is not evidence that they already had such knowledge by the time of the decision to reassign certain of Mole’s course lectures to other faculty members.
Where, as here, adverse employment actions or other problems with an employee predate any knowledge that the employee has engaged in protected activity, it is not permissible to draw the inference that subsequent adverse actions, taken after the employer acquires such knowledge, are motivated by retaliation. See, e.g., Clark County Sch. Dist. v. Breeden, 532 U.S. 268, 272 (2001) (per curiam) (employer commented about transferring plaintiff shortly prior to learning of plaintiff’s Title VII claim, and transferred plaintiff one month after learning of suit; employer “proceeding along lines previously contemplated, though not yet definitively determined, is no evidence whatever of causality”); Hoeppner v. Crotched Mountain Rehabilitation Ctr., 31 F.3d 9, 12, 14-16 (1st Cir. 1994) (employee already on probation prior to filing claim of sexual harassment; although discharge occurred within days of employee’s filing claim, summary judgment entered in favor of employer); Dziamba v. Warner & Stackpole LLP, 56 Mass. App. Ct. 397, 407 (2002) (summary judgment for employer on retaliatory discharge claim where “expressions of dissatisfaction, warnings about unsatisfactory accomplishment, and not raising his salary had begun well before” employee’s assertion of rights); Prader v. Leading Edge Prods., Inc., 39 Mass. App. Ct. 616, 617-618 (1996) (employee fired after receiving back pay award claimed retaliation; summary judgment for employer based on prior performance evaluation which, “although in general favorable to the plaintiff, stated that the plaintiff needed improvement” in various categories). Here, prior to the defendants’ awareness of Mole’s support of Anderson’s sexual harassment charge, there is undisputed evidence that his relationship with Czech had already been strained, that he was already losing financial support due to concerns of faculty members in various departments, and that his teaching assignments were already being *595transferred to other faculty. That evidence undercuts any inference that it was Mole’s subsequent support of Anderson’s charge that caused the various adverse employment decisions of which he now complains.
In addition to this defect in the sequence of events, the time span between Mole’s protected activity and the later adverse actions is too long to support his desired inference of causation. Where adverse employment actions follow close on the heels of protected activity, a causal relationship may be inferred. See Oliver v. Digital Equip. Corp., 846 F.2d 103, 110 (1st Cir. 1988). However, as the elapsed time between those two events becomes greater, the inference weakens and eventually collapses. “The cases that accept mere temporal proximity between an employer’s knowledge of protected activity and an adverse employment action as sufficient evidence of causality to establish a prima facie case uniformly hold that the temporal proximity must be ‘very close.’ ” Clark County Sch. Dist. v. Breeden, supra at 273, quoting O’Neal v. Ferguson Constr. Co., 237 F.3d 1248, 1253 (10th Cir. 2001). On a claim of retaliatory discharge, “unless the termination is very closely connected in time to the protected activity, the plaintiff must rely on additional evidence beyond temporal proximity to establish causation” (emphasis in original). Anderson v. Coors Brewing Co., 181 F.3d 1171, 1179 (10th Cir. 1999). See Figgous v. Allied/Bendix Corp., 906 F.2d 360, 362 (8th Cir 1990) (directed verdict for employer where only evidence was that employee had been fired one to two years after employer learned of his discrimination complaint); Oliver v. Digital Equip. Corp., supra at 110-111 (causation not shown where firing occurred more than two years after filing of charge); Cooper v. North Olmsted, 795 F.2d 1265, 1267, 1272 (6th Cir. 1986) (same; firing occurred four months after filing of charge). Here, Mole complains of acts that are many years removed from the employer’s discovery of his involvement in protected activity — he complains of salary reductions occurring three and four years, and of a termination that occurred nine years, after his support of Anderson’s charge became known. Mole cites no case suggesting that an inference of causation can be drawn when the adverse actions occur so many years after the protected activity.
*596Instead, Mole seeks to link those later acts to acts taken by Czech and Chlapowski closer in time to their becoming aware of Anderson’s charge. See Chungchi Che v. Massachusetts Bay Transp. Auth., 342 F.3d 31, 38 (1st Cir. 2003) (despite passage of eleven months between filing of claim and demotion of plaintiff, “[e]vidence of discriminatory or disparate treatment in the time period between the protected activity and the adverse employment action can be sufficient to show a causal connection”); Robinson v. Southeastern Pa. Transp. Auth., 982 F.2d 892, 895 (3d Cir. 1993) (ongoing “pattern of antagonism” sufficient to establish causal link between employee’s filing of grievance and his discharge two years later); Marx v. Schnuck Mkts., Inc., 76 F.3d 324, 329 (10th Cir.), cert. denied, 518 U.S. 1019 (1996) (lack of close temporal proximity not fatal to retaliatory discharge claim if “pattern of retaliatory conduct begins soon after [protected activity] and only culminates later in actual discharge”). In other words, a series of retaliatory measures starting shortly after the protected activity can justify the inference that a particular action in that series, although occurring a considerable time later, is still motivated by retaliation.
However, as discussed above with regard to the sequence of events, the “pattern” of adverse actions proffered by Mole was already underway prior to Czech and Chlapowski’s awareness that Mole had engaged in protected activity. Mole cannot invoke the concept of an ongoing pattern of actions against him and ignore when that pattern was established and how the events immediately following his protected activity fit into that pattern.15
The first act taken after Czech and Chlapowski knew of the *597pending sexual harassment charge was the October, 1990, annual evaluation of Mole. While that evaluation includes negative references to the fact that Mole’s teaching was curtailed, that event occurred at least four months earlier, prior to their awareness of any protected activity. And, to the extent the evaluation suggests anything negative about the status of funding for the PCF, it merely reflects actions that had been taken by the Scientific Council in February, 1990, prior to Anderson’s making any allegation of sexual harassment. The recommendation in that evaluation that Mole’s “primary focus and concern” should be to secure new sources of funding is not in any sense negative or adverse, and is, again, merely a continuation of a theme that had already been announced in his prior year’s evaluation (which had noted that his outside grant funding was coming to an end and that he should make “strong efforts to ensure continuation of his research funding”). In short, there is nothing adverse in that evaluation that was not already presaged by events occurring prior to Czech or Chlapowski’s awareness of Mole’s involvement in protected activity. This evaluation, the ostensible launching pad for Mole’s claim that every adverse action thereafter is linked to his protected activity, cannot be used to establish that the reductions in his salary (implemented three and four years later) and his ultimate termination (imposed nine years later) were caused by his engaging in protected activity. See Lewis v. Gillette Co., 22 F.3d 22, 25 (1st Cir. 1994) (claim of retaliatory discharge two years after participation in hearing not supported, where acts offered as causal link between those events were not themselves shown to be retalia*598tory; “there must be something more than a few weak inferences to create reasonable proof of a link between events so widely separated in time”).
While certain actions taken against Mole in 1991 and 1992 could be characterized as completely new adverse actions and not mere reiterations of concerns that had emerged prior to Mole’s engaging in protected activity, those acts are temporally removed from the discovery, in the late summer of 1990, that Mole had engaged in protected activity. We need not specify a precise outer time limit beyond which any inference of a causal link becomes impermissible. Suffice it to say that the time periods at issue in this case exceed the bounds that have been recognized by other courts. See, e.g., Mesnick v. General Elec. Co., 950 F.2d 816, 828 (1st Cir. 1991) (where some problems predated protected activity, and discharge occurred nine months after protected activity, sequence suggests absence of causal link; summary judgment entered in favor of employer); Cooper v. North Olmsted, supra (lapse of four months between protected activity and termination too long to support inference of causation).
The theory that these interim acts of Czech and Chlapowski create a causal link between Mole’s protected activity and his later salary reductions and termination also founders on the fact that those decisions were made by others, not by Czech or Chlapowski. Whatever may have motivated Czech and Chlapowski in the years following Anderson’s complaint, and even assuming that some of their actions in those years were motivated by retaliation, the acts that Mole ultimately complains about are acts taken by the evaluation committee and the chancellor of UMMC, not by Czech or Chlapowski.
Despite a retaliatory or discriminatory motive on the part of a supervisor who recommends that some adverse action be taken against an employee, a third person’s independent decision to take adverse action breaks the causal connection between the supervisor’s retaliatory or discriminatory animus and the adverse action. See Stimpson v. Tuscaloosa, 186 F.3d 1328, 1331-1332 (11th Cir. 1999), cert. denied, 529 U.S. 1053 (2000); Willis v. Marion County Auditor’s Office, 118 F.3d 542, 547-548 (7th Cir. 1997); Long v. Eastfield College, 88 F.3d 300, 307 (5th Cir. *5991996). The involvement of a third person as the actual decision maker will not break the causal connection if that decision maker merely “rubber stamps” the recommendation of the retaliating supervisor, or if the retaliating supervisor “dupes” the decision maker into taking action, or otherwise controls the decision maker. For example, in Cariglia v. Hertz Equip. Rental Corp., 363 F.3d 77, 87 & n.4 (1st Cir. 2004), a supervisor desirous of terminating an older worker because of his age gave a report to senior management omitting critical facts, thereby making it look as if that employee had misused company funds. Based on that report alone, without conducting any further investigation, and without even asking the employee for his version of events, management decided to fire the employee. Although there was no evidence that any of the senior managers harbored any bias against older workers, the supervisor’s own bias could still be the “cause” of the termination. Id. at 87-88. See Stimpson v. Tuscaloosa, supra at 1332 (biased recommender can still be cause of decision if “the decisionmaker followed the biased recommendation without independently investigating the complaint against the employee”); Long v. Eastfield College, supra at 307 & n.8 (if retaliating supervisor was college president’s sole source of information, and conduit through whom statements of other witnesses were funneled to president, president would no longer qualify as independent decision maker and termination decision would be tainted by supervisor’s retaliatory animus).
The mere fact that a retaliating supervisor provides some of the information on which a decision is based, or initially recommends the adverse employment action to someone higher up in the organization, does not necessarily mean that the decision maker lacks sufficient independence from the supervisor for these purposes. See Willis v. Marion County Auditor’s Office, supra at 547-548 (biased supervisor accused employee of tardy payment of invoices, with employee claiming that supervisor had planted false evidence; where higher management investigated accusations, and gave employee full opportunity to explain her side, causal connection between supervisor’s bias and termination was lacking); Long v. Eastfield College, supra at 307 (although retaliating supervisor recommended that two *600employees be fired, causal link between supervisor and president’s decision to terminate employees would be broken if president conducted his own investigation of allegations); Wilson v. Stroh Cos., 952 F.2d 942, 946 (6th Cir. 1992). See also Sattar v. Motorola, Inc., 138 F.3d 1164, 1171 (7th Cir. 1998) (biased supervisor not “some kind of Svengali” controlling management’s decision to terminate worker). When assessing the independence of the ultimate decision maker, courts place considerable emphasis on the decision maker’s giving the employee the opportunity to address the allegations in question, and on the decision maker’s awareness of the employee’s view that the underlying recommendation is motivated by bias or a desire to retaliate. See Cariglia v. Hertz Equip. Rental Corp., supra at 87 n.4; Stimpson v. Tuscaloosa, supra; Willis v. Marion County Auditor’s Office, supra at 547-548. Where those factors are present, they indicate that the decision maker has not merely accepted the tainted recommendation at face value, but has in fact made a sufficiently independent determination as to whether discipline or adverse action is appropriate.
Here, all that Mole can show is that Czech and Chlapowski gave negative evaluations of his work and made adverse recommendations against him, which were seen by the evaluating committee when they performed evaluations of Mole in subsequent years, and that Czech and Chlapowski provided negative information to the chancellor in connection with the chancellor’s proceedings to terminate him. It is undisputed, however, that Czech and Chlapowski were not the sole sources of information for either the committee or the chancellor; that each of the decision makers was made aware of Mole’s view that Czech and Chlapowski were retaliating against him; and that each of the decision makers gave Mole the opportunity to present his version of events and his explanations for his conduct. Indeed, the members of the evaluation committee knew that the very reason for its existence was a grievance concerning allegations of unfair evaluations by Chlapowski. They had been selected, by vote of the department faculty, as the persons to take over the task of conducting Mole’s evaluation for the express purpose of providing an evaluation independent from *601Chlapowski.16 Their first evaluation expressly addressed Mole’s allegation that he had been “unfairly denied the opportunity to teach and serve on committees,” concluding that there were sound reasons for both of those decisions. Mole was given the opportunity to submit, both in writing and orally, his own views on the matter of his evaluation. Moreover, each of those evaluations addressed Mole’s performance in later time periods, considering new matters that had never even been part of prior evaluations by Czech or Chlapowski (e.g., their assessment that Mole had deliberately sabotaged a 1995 grant application). Nothing in the evidence suggests that Czech or Chlapowski was “the de facto decisionmaker” behind the committee’s evaluations or that the committee was “duped into being the conduit for their retaliation.” Willis v. Marion County Auditor’s Office, supra at 548.
The same is true with respect to the chancellor’s decision to recommend Mole’s termination. The chancellor’s information was not limited to that provided by Czech or Chlapowski. He had information from the evaluation committee (which had by then abandoned its work as hopeless), and the views of other faculty members (including the many members of the department who had been seeking Mole’s resignation or termination as far back as 1993).17 During the lengthy formal proceedings to consider the termination of a tenured faculty member, Mole (represented by counsel) had a full opportunity to be heard on *602the criticisms that were being leveled against him from various quarters, and to present his claim that what was being portrayed as his own failure was attributable to unfair treatment at the hands of Czech and Chlapowski. At trial, he presented no evidence to suggest that the chancellor was motivated by a desire to retaliate against him for protected activity, and no evidence that the chancellor, at the end of those lengthy proceedings, was merely “duped into being the conduit” for Czech or Chlapowski’s retaliation. Id.
Viewed in the light most favorable to Mole, the evidence was that his various problems at UMMC predated his involvement in any protected activity; that the adverse actions complained of were taken years after his supervisors became aware of his protected activity; and that those adverse actions were taken by faculty members and university officials independent of the supervisors who allegedly harbored a desire to retaliate against him for that protected activity. The issue before us is not whether Mole was treated fairly with respect to the numerous setbacks he encountered during his final decade at UMMC. The issue is whether he has produced sufficient evidence to show that he suffered those setbacks because of his endorsement or encouragement of his wife’s decision to lodge a charge of sexual harassment. He has not done so, and the defendants were therefore entitled to a directed verdict in their favor on the claims of retaliation.
b. Section 1983. Mole also brought a claim under 42 U.S.C. § 1983 against Czech and Chlapowski, alleging that they retaliated against him because of his exercise of his right of intimate association with his wife. See Sowards v. Loudon County, 203 F.3d 426, 431-434 (6th Cir.), cert. denied, 531 U.S. 875 (2000); Adler v. Pataki, 185 F.3d 35, 41-44 (2d Cir. 1999); Adkins v. Board of Educ. of Magoffin County, 982 F.2d 952, 955-956 (6th Cir. 1993). For purposes of such a claim, the plaintiff must prove that “(1) the plaintiff engaged in protected conduct; (2) an adverse action was taken against the plaintiff that would deter a person of ordinary firmness from continuing to engage in that conduct; and (3) there is a causal connection between elements one and two — that is, the adverse action was motivated at least in part by the plaintiffs protected conduct.” *603Sowards v. Loudon County, supra at 431, quoting Thaddeus-X v. Blatter, 175 F.3d 378, 394 (6th Cir. 1999) (en banc). For this claim, the causation analysis focuses on whether the adverse action “was motivated in substantial part by the plaintiff’s constitutionally protected activity.” Sowards v. Loudon County, supra, citing Mattox v. Forest Park, 183 F.3d 515, 520-521 (6th Cir. 1999). Thus, Mole was again required to show a causal connection between his protected conduct (his exercise of a constitutional right of association with Anderson) and the adverse actions taken against him. For the reasons discussed above, the evidence was insufficient to show such a causal connection. Recasting the protected activity as the exercise of a right of association does not alter the analysis of the causation element of his claim.18,19
Judgment affirmed.
Directed verdicts in favor of Czech and Bratt were entered at the close of the plaintiff’s case. Directed verdicts in favor of Chlapowski and the University of Massachusetts were entered at the close of all the evidence. On appeal, the plaintiff makes no argument concerning the directed verdict in favor of Bratt, and we therefore limit our analysis to the claims involving the other three defendants.
Czech and Chlapowski were not members of the Scientific Council, and neither of them had served on the subcommittee.
The precise date of the decision to cut back on Mole’s teaching is unclear, but it is referenced in a letter from Mole to other faculty members dated June 6, 1990, thereby establishing that the decision had been made and communicated to Mole on or before that date.
There is evidence that Czech recommended that DERC take that action, but no evidence that Czech participated in the vote.
Although the Scientific Council’s subcommittee had recommended the cessation of funding over a year earlier, it appears that that recommendation had not been acted on formally. However, the evidence suggests that funding had ceased as a practical matter in the wake of the subcommittee’s report.
The university had notified Anderson that her contract would not be renewed after the 1991-1992 academic year. That notification had in turn triggered a dispute when Chlapowski attempted to withhold the department’s twenty per cent share of Anderson’s salary and transfer responsibility for her entire salary to Mole. Mole successfully pursued a grievance on the allocation of responsibility for Anderson’s salary during the final year of her contract.
Mole filed a grievance with respect to the location of his laboratory, but was unsuccessful in blocking the move. The space to which he was moved was smaller than his prior space, but was comparable to the amount of laboratory space allotted to other researchers in the department. Despite this move, the space formerly used by Mole was not reassigned to another researcher, but was used for miscellaneous storage.
At trial, two faculty members testified to statements by Mole to the effect that he had no intention of working but that the department would still have to support him. While the jury could choose not to credit that testimony, it is undisputed that nine faculty members wrote to the chancellor expressing the view that Mole was being “deliberately unproductive.”
Mole had not followed the instructions for submitting the grant application and the application had therefore been returned promptly to him. He nevertheless had characterized the application as “pending.”
Specifically, in two places on the application, Mole had stated that the department had not given him sufficient laboratory space to do the research outlined in the grant application, and had drawn attention to this statement by way of underlining. In the evaluation committee’s view, “submission of an application with the inclusion of such statements cannot be considered as a serious effort to obtain funding.”
Mole contends that the chancellor’s decision to commence those proceedings was triggered by Chlapowski’s sending a letter to Edward Bresnick (a vice-chancellor for research at UMMC). The evidence was that Bresnick had *591solicited information from Chlapowski, and that Chlapowski’s letter came in response. There is no evidence with respect to what role that letter played (if any) in the chancellor’s decision to seek Mole’s termination the following year. The chancellor testified that the decision to take action was prompted by the resignation of the evaluation committee and by the difficulties he was having in recruiting anyone to take over as chair of the department. (A new acting chair, Anthony Carruthers, had taken over in 1997, on the condition that he not be required to have any role in evaluating Mole.)
For purposes of a claim under G. L. c. 151B, § 4 (4), a plaintiff has engaged in protected activity if “he has opposed any practices forbidden under this chapter or . . . has filed a complaint, testified or assisted in any proceeding under [G. L. c. 151B, § 5].” Under § 4 (4A), protected activity consists of “aid[ing] or encouraging] any other person in the exercise or enjoyment of” any right under G. L. c. 151B. With respect to a retaliation claim under Title VII, a person has engaged in protected activity if “he has opposed any practice made an unlawful employment practice by [Title VII], or . . . has made a charge, testified, assisted, or participated in any manner in an investgation proceeding, or hearing under [Title VH].” 42 U.S.C. § 2000e-3(a). Here, Anderson’s complaint of sexual harassment was ultimately unsuccessful, as a jury rendered verdicts against her, and those judgments were affirmed on appeal. Anderson-Mole v. University of Mass., 49 Mass. App. Ct. 723 (2000). The fact that a complaint is later found to be unmeritorious does not preclude a retaliation claim based on the protected activity of pursuing *592that complaint. See Mesnick v. General Elec. Co., 950 F.2d 816, 827 (1st Cir. 1991), cert. denied, 504 U.S. 985 (1992). The defendants make no argument concerning the evidence that Mole engaged in protected activity.
Under G. L. c. 15IB, § 4 (4), adverse actions consist of a defendant’s action “to discharge, expel or otherwise discriminate against” the plaintiff. Under § 4 (4A), adverse action is any action “to coerce, intimidate, threaten, or interfere with” the plaintiff. For a Title VII claim, the adverse action consists of an employer’s decision to “discriminate against” the plaintiff. 42 U.S.C. § 2000e-3(a). The defendants make no argument with respect to the evidence that Mole suffered adverse action.
Mole’s claim with regard to an alleged series of retaliatory actions that only later resulted in salary reduction and termination is further complicated by the fact that a retaliation claim for those earlier acts would be time-barred. Mole did not file his claim until May 2, 1993. At the time, the limitations period established by G. L. c. 151B, § 5, as amended through St. 1989, c. 722, §§ 24-29, was six months, thus limiting Mole’s G. L. c. 151B retaliation claim to acts committed on or after November 2, 1992. (That limitations period was later extended to 300 days. St. 2002, c. 223, § 1.) The limitations period for Mole’s Title VII claim was 240 days, 42 U.S.C. § 2000e-5(e) (2000), thus limiting his Title VII claim to acts occurring on or after September 2, 1992. See 29 C.F.R. § 1601.13(b) (2004). On the defendants’ summary judgment motion, a judge in the Superior Court ruled that the theory of *597“continuing violation” did not apply, and allowed partial summary judgment in favor of all defendants with respect to the allegedly retaliatory acts that occurred prior to those dates. Mole has not appealed from that ruling.
Mole contends, however, that those time-barred events were nevertheless admissible as “background evidence,” Cuddyer v. Stop & Shop Supermarket Co., 434 Mass. 521, 530 n.10 (2001), citing Sabree v. United Bhd. of Carpenters & Joiners, Local No. 33, 921 F.2d 396, 400 n.9 (1st Cir. 1990), and that that “background” may be used to prove a causal link between his engaging in protected activity and the adverse actions taken years later. The use of time-barred acts to fill in a lengthy gap between the protected activity and the subsequent acts complained of is a “vexing question.” Costello v. Massachusetts Rehabilitation Comm’n, 982 F. Supp. 61, 67 n.9 (D. Mass. 1997). We need not resolve that “vexing question” in this case, because the inference of causation suggested by Mole is impermissible for other reasons.
Mole’s assertion that that committee also was biased against him, and that he had not exercised his option to select his own member of the evaluation team because the entire department was biased against him, does not help his cause. He has no evidence to suggest that the reason for any other faculty member’s animosity toward him was due to his involvement with his wife’s complaint. Whether the members of the evaluation committee had their own reasons for disliking Mole, they were fully aware of the fact that they had been convened to perform their independent evaluation of Mole’s work and performance, not merely to implement recommendations contained in earlier evaluations.
The chancellor’s testimony characterized these sources of information as “a cascade of misery from all directions.” Evidence of complaints from or acrimonious relations with colleagues other than the alleged retaliators (which is undisputed on this record) has sufficed to defeat the causation prong of a prima facie case. See McMillan v. Massachusetts Soc’y for the Prevention of Cruelty to Animals, 140 F.3d 288, 309 (1st Cir. 1998), cert. denied, 525 U.S. 1104 (1999).
In a single sentence relegated to a footnote, Mole contends that the conduct at issue for this claim “is not restricted to post-September, 1992, actions by the applicable limitations period, which is much longer than those controlling the discrimination statutes,” citing Piacentini v. Levangie, 998 F. Supp. 86, 88 (D. Mass. 1998). See G. L. c. 260, § 5B (actions alleging violations of civil rights to be brought within three years from date cause of action accrues). We do not consider arguments raised only in footnotes, see Commonwealth v. Clerk-Magistrate of the W. Roxbury Div. of the Dist. Court Dep’t, 439 Mass. 352, 361 n.7 (2003), or arguments raised in but a single sentence, see Commonwealth v. Donahue, 430 Mass. 710, 714-715 n.1 (2000). The single-sentence footnote in Mole’s brief does not identify when his action was filed, his theory of when it accrued, or which additional acts of Czech or Chlapowski would then come within the “longer” limitations period. His footnote thus does not address how the inclusion of those unspecified additional acts might strengthen the evidence of causation.
The defendants argue that there were other defects in Mole’s § 1983 claim. Because there was insufficient evidence of a causal connection between the alleged protected activity and the adverse actions complained of, we need not address those other arguments.